As outlined in the Case Information Catchwords on the High Court’s webpage, the Special Case in Zhang v Commissioner of Police appeared to raise a bundle of interesting questions about the operation of the implied freedom of political communication in a global context, and the extent of police powers concerning foreign interference offences under Commonwealth legislation. However, when judgment was delivered, it appeared to be a bit of a fizzer – none of the interesting questions about the implied freedom of political communication were touched upon in the High Court’s unanimous decision, with the Court ultimately deeming the questions in the Special Case “unnecessary to answer”.

Nevertheless, although the decision does not assist in our understanding of the implied freedom of political communication, it remains an important judgment for constitutional lawyers and academics. It is an important illustration of the principle that a statute should, where possible, be read so as to avoid constitutional invalidity. The decision also demonstrates some of the risks of the Special Case procedure for the determination of constitutional questions.

The Facts

The facts giving rise to the proceedings were particularly interesting – the plaintiff (Zhang) was an Australian citizen, born in the People’s Republic of China, who was employed in the office of a (later suspended) New South Wales Member of Parliament. For some time, Zhang had been under investigation by the Australian Federal Police (AFP) in relation to suspected offences under s 92.3(1) and (2) of the Commonwealth Criminal Code. These “reckless foreign interference” offences criminalise, in general terms, conduct engaged in on behalf of, or with the collaboration of, a foreign government, where the person is reckless as to whether the conduct will, inter alia, influence a political or governmental process, support a foreign intelligence activity or (in the case of an offence against s 92.3(1)), prejudice national security. (The differences between the two provisions are subtle, s 92.3(2) is directed towards the influence upon a person in relation to a particular process, or exercise of a right or duty, whereas s 92.3(1) is directed towards the influence on that process, or exercise of a right or duty.)

During the course of the investigation, the AFP obtained search warrants under the Crimes Act 1914 (Cth). Those warrants purported to authorise search and seizure of material relevant to the suspected foreign interference offences. When the warrants were executed, the AFP seized material (including mobile phones and computers) relevant to the foreign interference offences (as well as other material that the AFP officers believed was relevant to offences under other provisions of the Commonwealth Crimes Act).

Zhang commenced proceedings in the original jurisdiction of the High Court, seeking declarations that s 92.3(1) and (2) of the Criminal Code were invalid, together with writs of certiorari quashing the warrants and a mandatory injunction requiring the destruction or return of the seized and copied material. By a Special Case, Zhang and the Commissioner for Police agreed to the statement of various questions of law.

In that Special Case, Zhang challenged the validity of each warrant on two grounds. First, he contended that each warrant was invalid because it did not identify the substance of the offences with sufficient precision. Second (and more interestingly), he contended that each warrant failed to authorise the search and seizure because s 92.3(1) and (2) each contravened the implied freedom of political communication.

In contending that s 92.3(1) and (2) were invalid, Zhang observed that, for each of the offences created by s 92.3, the effect of the criminalised conduct related to the influence of political or governmental processes and/or democratic and political rights and duties. Zhang argued that “criminalisation of conduct which may have the effect of influencing a person in that regard, or the particular process, right or duty, limits the making or content of political communications and thus burdens the implied freedom”: Zhang’s written submissions at [33].

Zhang further submitted that, as the legislation burdened political communication, it was therefore necessary to address the legitimacy of the purpose of the legislation. In this respect, Zhang noted that whilst the extrinsic materials indicated that the purpose of the provision was to avoid harm to Australian interests, such harm was not an element of an offence under either provision. Rather, the minimum quality of the conduct necessary for the offence was that the conduct be “covert”: at [37]. By reference to its ordinary meaning (which stood in contrast to terms such as “deceptive” or “clandestine”) and the extrinsic materials, Zhang argued that the term “covert” should be read as including any conduct which is not openly acknowledged, shown or engaged in, or which lacks transparency or is private”: at [37]. In particular, this would include any communication on private social media.

Whilst acknowledging that foreign interference may have “serious consequences for liberal democratic systems of government”, and that undisclosed or non-transparent foreign influence can have “serious implications for sovereignty and national policy as it may result in the prioritisation of foreign interests over domestic interests”, Zhang contended that the breadth of the provisions overreached any legitimate purpose, such that the provisions were not proportionate to the potential threat: at [49].

In response, the Commissioner of Police argued that “properly construed, the provisions have a narrower operation than that for which the plaintiff contends”: Commissioner’s written submissions at [6.2]. The Commissioner emphasised that the legislation only criminalised “covert” communications, and submitted that this required the prosecution to establish that the accused took “action to conceal, hide, keep secret or disguise the relevant conduct”: at [21]. The Commissioner submitted that, so construed, the legislation burdened the implied freedom, at most, to a very limited extent, and that the legislation was reasonably appropriate and adapted to a legitimate purpose, namely the protection of Australia’s sovereignty, by reducing the risk of foreign interference in Australia’s political or governmental processes: at [26], [33] and [39] – [44].

During the course of oral argument, Gageler J squarely raised the significance of this question of statutory construction. After observing that the “real thrust” of Zhang’s argument was “one of which the Americans would call ‘overbreadth’”, Gageler J put the following proposition to senior counsel for Zhang:

GAGELER J: Without being too simplistic, I hope, if you read ‘covert’ broadly, you say you win; if you read ‘covert’ narrowly, then there is no overbreadth problem.

Senior counsel accepted this characterisation of the submissions, responding:

MR WALKER: That is a useful simplification, with respect. I do not intend by my assent to it to dispense with everything I have advanced in support of that, but yes, that is the way in which that part of the argument proceeds.

Previously in the hearing, Edelman J had had the following exchange with senior counsel:

EDELMAN J: If ‘covert’ did have those broad connotations, why could it not be read down to have the narrower meaning?

MR WALKER: In a criminal case one would no doubt argue that, in my position, yes…

The Full Court of the High Court unanimously held that this concession was fatal to Zhang’s constitutional contentions. At [20], the Court explained:

Implicit in Mr Zhang’s failure to assert that the word ‘covert’ in s 92.3(1)(d)(i) would be incapable of being read down to ensure validity was an acknowledgement that those parts of s 92.3(1) which support the offences against s 92.3(1) to which each warrant relates (being s 92.3(1)(a) read with the first part of s 92.3(1)(b)(i), s 92.3(1)(c)(i) and (ii) and the first part of s 92.3(1)(d)(i)) have some valid operation. That being so, his argument that those offences do not exist can be rejected without need of determining the constitutional argument he presents and without need of determining the attendant question of the proper construction of the word.

The Court accordingly answered the questions in the Special Case relating to the validity of s 92.3 as unnecessary and inappropriate to answer.

A Fizzer?

For anyone anxiously awaiting the outcome in Zhang for its discussion of the implied freedom of political communication and its relationship to foreign interference offences in a world of instant global communication, the case was definitely a fizzer. The High Court did not engage with any of the constitutional contentions, and did not touch upon any of the interesting issues that were addressed in the parties’ submissions.

The High Court’s decision in Zhang should nonetheless remain of interest to constitutional lawyers. In particular, Zhang is a particularly strong illustration of the role of statutory interpretation in constitutional law, and of the importance of the principle that a statute should, where possible, be read so as to avoid constitutional invalidity.

In this respect, Zhang is another example of a case in which the arguments “divid[ed] along battlelines” where “the party seeking to challenge validity advance[d] a literal and draconian construction, even though the construction would be detrimental to that party were the law to be held valid”, whereas “the party seeking to support validity advance[d] a strained but benign construction, even though the construction is less efficacious from the perspective of that party than the literal construction embraced by the challenger”: Zhang at [27], citing Northern Australian Aboriginal Justice Agency Ltd v Northern Territory [2015] HCA 41 (NAAJA v NT) at [75], per Gageler J.

As the Court in Zhang observed, this subversion of the parties’ positions also occurred in NAAJA v NT. In those proceedings, NAAJA contended for a wide construction of the ‘paperless arrest laws’, such that the amending provisions would have authorised a person to be taken into custody for an infringement notice offence and detained for up to four hours; whilst the NT government’s position was that detention was only authorised for so long as was reasonable for a decision to be made as to what action should be taken in respect of that person (an interpretation that did not accord with the purpose of the provisions as explained in the extrinsic material: NAAJA v NT at [88], per Gageler J).

Such a reversal of positions was also seen, albeit to a lesser extent, in Minister for Home Affairs v Benbrika [2021] HCA 4, where Mr Benbrika argued that legislation empowering a Supreme Court to order a person convicted of certain terrorist offences to be detained following the expiration of his or her sentence was unconstitutional as authorising detention of a citizen for a punitive purpose. In response, the Commonwealth Attorney-General emphasised the limits of the legislation, pointing out that a continuing detention order could only be made where the offender poses an unacceptable risk of committing a serious terrorist offence and “there is no other less restrictive measure that would be effective in preventing the unacceptable risk”: Commonwealth Attorney-General’s submissions at [52]. Indeed, it may be observed that one of the significant differences between the decisions of the majority and the dissenting Judges in Benbrika is the extent to which the Judges accepted the significance of this limitation emphasised by the Commonwealth: Benbrika at [47], per Kiefel CJ, Bell, Keane and Stewart JJ; at [100] – [101], per Gageler J; at [170], per Gordon J; and at [236], per Edelman J.

As Gageler J observed in NAAJA v NT at [75], such reversals in the parties’ usual positions are “not unfamiliar where questions about the constitutional validity of a law are abstracted from questions about the concrete application of that law to determine the rights and liabilities of the parties”.

It may be noted that each of Benbrika, NAAJA v NT and Zhang were cases in which abstract questions were posed to the Court for determination prior to the finding of any fact (either by removal of a question reserved or through a Special Case). In this respect, Zhang illustrates the risks of this form of such Stated Case procedures for the determination of constitutional questions. By permitting the abstraction of constitutional issues at a point in the litigation where no determination has yet been made about the application of the statute to the particular factual situation, the procedure channels the parties into reversed positions on questions of statutory construction: the government is driven to read the legislation down in order to maximise the prospect that the legislation will be found to be valid, whereas the challenging party is driven to adopt an overbroad construction, so as to maximise the prospect that the legislation will be found to be unconstitutional.

As the United States Supreme Court has observed (in the context of what is referred to in the US as a “facial challenge” to legislation), such an approach risks the “premature interpretation of statutes on the basis of factually barebones records”: Sabri v United States, 541 U.S. 600, 609 (2004); see similarly Tajjour v New South Wales (2014) 254 CLR 508 at [175], per Gageler J (referring to the problems that arise where constitutional adjudication is conducted in a “factual vacuum”). Such a case will lack “that clear concreteness provided when a question emerges precisely framed and necessary for decision from a clash of adversary argument exploring every aspect of a multi-faced situation embracing conflicting and demanding interests”: Zhang at [25], citing Mellifont v Attorney-General (Qld) (1991) 173 CLR 289 at [11] and Kuczborski v Queensland (2014) 254 CLR 51 at [186].

The Court in Zhang concluded (citing Gageler J in NAAJA) that “a court should be wary” where this reversal occurs. The Australian adversarial system proceeds on the basis that legal questions are best determined when two parties with a direct interest in the proceeding fully and completely present their case before an independent tribunal. When constitutional questions are determined prior to the determination of the underlying facts and prior to the application of the statute to those facts, the determination of the constitutional issues will tend to become hypothetical in nature. The determination of such hypothetical questions “run[s] counter to the administration of justice in an adversary system”: Zhang at [25].


The decision in Zhang clearly illustrates the risks of Stated Case procedures for the determination of constitutional questions. When constitutional questions are posited prior to the determination of factual questions, it will not be uncommon for the parties’ positions to be reversed on underlying issues of statutory interpretation, with the government arguing for a (potentially) unduly narrow construction of the legislation and the challenging party arguing for a (potentially) unduly broad construction of the legislation. Such a reversal of positions potentially undermines the adversarial system, and risks questions of constitutional validity being determined without full consideration of the nuanced and varied ways in which the legislation may be applied (or misapplied) in a particular situation. For these reasons, as the Court held in Zhang, courts “should be wary” whenever constitutional questions are posed for consideration prior to the determination of the relevant facts in issue between the parties.

Belinda Baker is a fellow of the Gilbert + Tobin Centre of Public Law at UNSW Law & Justice and a Deputy Senior Crown Prosecutor at the NSW ODPP.

Suggested citation: Belinda Baker, ‘“A Court should be Wary” Zhang v Commissioner of Police [2021] HCA 16′ on AUSPUBLAW (16 June 2021) <>