• Specific Year
    Any

Pender, Kieran --- "Comcare v Banerji: Public Servants and Political Communication" [2019] SydLawRw 5; (2019) 41(1) Sydney Law Review 131


Before the High Court

Comcare v Banerji: Public Servants and Political Communication

Kieran Pender[*]

Abstract

In March 2019 the High Court of Australia will, for the first time, consider the constitutionality of limitations on the political expression of public servants. Comcare v Banerji will shape the Commonwealth of Australia’s regulation of its 240 000 public servants and indirectly impact state and local government employees, cumulatively constituting 16 per cent of the Australian workforce. But the litigation’s importance goes beyond its substantive outcome. In Comcare v Banerji, the High Court must determine the appropriate methodology to apply when considering the implied freedom of political communication’s operation on administrative decisions. The approach it adopts could have a significant impact on the continuing development of implied freedom jurisprudence, as well as the political expression of public servants.

I Introduction

Australian public servants have long endured an ‘obligatiopn of silence’.[1] Colonial civil servants were subject to strict limitations on their ability to engage in political life.[2] Following Federation, employees of the new Commonwealth of Australia were not permitted to ‘discuss or in any way promote political movements’.[3] While the more draconian of these restrictions have been gradually eased, limitations remain on the political expression of public servants. Until now, these have received surprisingly little judicial scrutiny. Although one of the few judgments in this field invalidated the impugned regulation,[4] the Australian Public Service (‘APS’) has continued to limit the speech of its employees.

These limitations give rise to tension between competing policy interests.

On the one hand, there is significant public interest in an impartial and apolitical bureaucracy.[5] In Australia’s Westminster/Whitehall model, the politicisation of the public service would have considerable adverse consequences. The ‘Government of the day’ must have confidence in the quality of the public service ‘irrespective of which political party is in power’,[6] to prevent ‘the insecurity and ineptitude of a reversion to political patronage’.[7] The Commonwealth, as an employer, is also entitled to expect that its employees obey certain contractual obligations (such as the duty of fidelity), just as a private sector employer would not tolerate overt criticism from an employee.[8] But, on the other hand, the wholesale exclusion of government employees from political debate has deleterious effects. A complete prohibition would quantitatively degrade political discourse given the size of the restricted class,[9] and have a qualitative impact given public servants are often ‘uniquely qualified to comment’ on policy matters.[10] ‘Indeed it would be inappropriate’, the Commonwealth Public Service Board once admitted, ‘to deprive the political process of the talent, expertise and experience of individuals simply because they are employed in the public sector.’[11] There are also rights-based concerns: ‘because we have not relegated our officials to the status of second class citizens’, public servants have a reasonable expectation of political enfranchisement.[12] Appropriately balancing these interests in a constitutional democracy is no easy task.

Comcare v Banerji provides the High Court of Australia with an opportunity to consider this tension and how restrictions on public servants’ political expression, first developed in the mid-1800s, interact with two more contemporary developments: the implied freedom of political communication in the Australian Constitution and social media. Three primary issues arise from the Commonwealth’s termination of a public servant’s employment in relation to her use of the social media platform Twitter and her subsequent workers’ compensation claim. First, how should the relevant limitations on political expression contained within the Public Service Act 1999 (Cth) (‘Public Service Act’) be interpreted? Second, what is the appropriate methodology for reviewing the constitutionality of the termination? Third, is the termination consistent with the implied freedom of political communication? These questions are not entirely conceptually distinct and there is some overlap between them. However, for the purposes of clarity, this article will consider them each in turn after providing a brief background to Comcare v Banerji.

II The Facts and Litigation

In May 2011,[13] Ms Michaela Banerji joined Twitter. Banerji, a public affairs officer with what was then called the Department of Immigration and Border Protection (‘the Department’), elected to tweet under the pseudonym ‘LaLegale’. She proceeded to tweet frequently, often criticising the Federal Government, relevant ministers and bureaucrats in relation to border protection policy. The primary sentiment of her tweets was that Australia’s treatment of asylum seekers was ‘unlawful, immoral and destructive’.[14] Her tweets did not disclose any confidential information. With the exception of one tweet, her comments were made outside of work hours and exclusively using personal communication devices.[15]

In March 2012, a fellow employee of the Department complained that Banerji’s use of social media was in breach of the APS Code of Conduct, contained within s 13 of the Public Service Act. At the time, s 13(11) required APS employees to behave ‘at all times’ in a manner that ‘upholds the APS Values and the integrity and good reputation of the APS’. Among the APS Values articulated at the time in s 10(1) was that the ‘APS is apolitical, performing its functions in an impartial and professional manner’.[16] Initially, the Department determined that there was insufficient evidence to proceed with an investigation. However, in May 2012, upon receipt of additional information from the complainant, the Department proceeded to investigate Banerji’s suspected contravention of s 13(11). In September 2012, Banerji was advised that the Department proposed to make a finding that she had contravened the APS Code of Conduct. Section 15 provided a discretionary power for the Department to impose various disciplinary measures, including termination of Banerji’s employment.

A Banerji v Bowles (Federal Circuit Court of Australia)

In October 2012, Banerji filed a general protections application with Fair Work Australia. She simultaneously commenced proceedings in the Federal Circuit Court of Australia seeking an interlocutory injunction to prevent the Department from terminating her employment.[17] Banerji grounded her application in a broad conception of the implied freedom of political communication. She placed reliance on comments of Kirby J in Australian Broadcasting Corporation v Lenah Game Meats[18] that, she argued, recognised a broad ‘right to express political opinion’.[19]

Judge Neville was unwilling to accept this radical departure from implied freedom orthodoxy. In a judgment delivered in August 2013, his Honour refused to issue an injunction: ‘The unbridled right championed by Ms Banerji ... does not exist.’[20] Although deferring any substantive hearing on the constitutional claim to a superior court, Judge Neville added some further observations. His Honour said:

I do not see that Ms Banerji’s political comments, ‘tweeted’ while she remains (a) employed by the Department, (b) under a contract of employment, (c) formally constrained by the APS Code of Conduct, and (d) subject to departmental social media guidelines, are constitutionally protected.[21]

In September 2013, the Department advised Banerji that her employment had been terminated pursuant to s 29(1) of the Public Service Act.[22]

B Banerji v Comcare (Administrative Appeals Tribunal)

Following the termination of her employment, Banerji suffered from an adjustment disorder characterised by depression and anxiety. Accordingly, in October 2013 she lodged a claim for workers’ compensation under the Safety, Rehabilitation and Compensation Act 1988 (Cth). In February 2014, a Comcare delegate refused Banerji’s claim. She subsequently requested a reconsideration, and the refusal was affirmed. In September 2014, Banerji sought merits review of the decision in the Administrative Appeals Tribunal (‘the Tribunal’). The primary matter in dispute was whether the Department’s termination of Banerji’s employment constituted reasonable administrative action taken in a reasonable manner. If so, her claim was destined to fail. However, if — as Banerji contended — the termination was unlawful and therefore could not constitute reasonable administrative action, she would be entitled to workers’ compensation.

At this juncture, it is necessary to highlight the current contours of the implied freedom of political communication. First outlined by the High Court in 1992, the freedom took on a more settled form in 1997 in Lange v Australian Broadcasting Corporation.[23] It has subsequently been modified by several cases, including Coleman v Power,[24] McCloy v New South Wales,[25] and Brown v Tasmania.[26] Its most recent authoritative formulation, by Kiefel CJ, Bell and Keane JJ in Brown, asks,

in relation to an impugned law:

1. Does the law effectively burden the freedom in its terms, operation or effect?

2. If ‘yes’ to 1, is the purpose of the law legitimate, in the sense that it is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government?

3. If ‘yes’ to 2, is the law reasonably appropriate and adapted to advance that legitimate object in a manner that is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government? This question involves ‘proportionality testing’ to determine whether the restriction that the provision imposes on the freedom is justified. The proportionality testing involves three stages, being inquiries as to whether the law is justified as suitable, necessary and adequate in its balance.[27]

Against this context, the Tribunal’s Deputy President Humphries and Member Hughson found for Banerji.[28] After determining that s 13(11) of the APS Code of Conduct burdened the freedom of political communication, the Tribunal applied the Lange test. The Tribunal firstly accepted Comcare’s submission that ‘maintaining an apolitical public service, and maintaining public confidence in that service’ was consistent with the constitutionally-prescribed system of government.[29] After outlining the approach to proportionality testing articulated in McCloy, the Tribunal observed that if Banerji’s tweets had been attributable, the imposition of sanctions ‘would have constituted a proportionate and appropriate application of a law competently designed to preserve the [APS’s] apolitical and impartial status’.[30]

Yet according to the Tribunal, Banerji’s comments made under a pseudonym were not attributable — to her personally or public servants as a class. The Tribunal placed considerable importance on this distinction. ‘The explicit objectives of a law designed to protect the impartial status of the APS,’ the Tribunal continued, ‘fall away in the context of comments not ostensibly made by a public servant.’[31] Within the McCloy proportionality analysis, the Tribunal balanced ‘a serious impingement on Ms Banerji’s implied freedom’ with ‘a law only weakly and imperfectly serving a legitimate public interest’.[32] The anonymous character of the comments meant that ‘the balance tips markedly in Ms Banerji’s favour’.[33] Accordingly, the Tribunal held that the termination of Banerji’s employment ‘unacceptably trespassed on the implied freedom of political communication’, such that the termination could not constitute reasonable administrative action.[34]

In reaching its decision, the Tribunal addressed, but did not resolve, a methodological question that has assumed some importance in the appeal. The Tribunal had proceeded on the basis that Banerji’s challenge to the termination required an analysis of the validity of the statute itself. While the Tribunal lacked jurisdiction to invalidate s 13(11), its finding that the termination was ultra vires was made on the basis that the legislation contravened Lange/McCloy. An alternative view is that the statute is to be construed so that it does not authorise an exercise of power that impermissibly infringes on the freedom, with the result that the actual exercise of the power — terminating Banerji’s employment — is ultra vires for exceeding statutory authority. This would be to adopt an administrative, rather than constitutional, review to constrain the exercise of power. Although the Tribunal noted it was ‘not the present task’ to determine the methodological issue,[35] it did reason that s 13(11) ‘evidently empowers’ the termination of Banerji’s employment in such circumstances, and as such ‘it is the empowering statute which placed the burden on political communication, and not the act of the delegate’.[36] The Tribunal observed:

The words at all times must be given their ordinary meaning ... those words have the effect of extending the temporal operation of the Code ... [and importing] the notion that the values must be upheld whatever ‘hat’ the employee was wearing.[37]

The Tribunal observed that reading down or severing the offending part of the Public Service Act ‘would be no easy matter’.[38]

C Comcare v Banerji (High Court of Australia)

Comcare appealed the Tribunal’s decision, and in September 2018 the dispute was removed to the High Court. The substantive submissions in support of Comcare’s notice of appeal are those of the Attorney-General of the Commonwealth, intervening (‘the Commonwealth’). The Commonwealth submits that the termination of Banerji’s employment was reasonable administrative action and her Comcare claim must fail.[39] The Commonwealth makes two primary, alternative contentions: (A) s 13(11) is valid in all of its possible applications;[40] or (B) to the extent that the individual exercise of power becomes relevant, in this case it was exercised consistently with the implied freedom.[41]

Banerji’s position is also multi-layered. If the Court does not accept her primary submission (A) that on its correct interpretation s 13(11) did not extend to her conduct, she submits (B) that the termination itself is invalid. Banerji argues that the decision-maker failed to consider a mandatory relevant consideration (the implied freedom), and the decision itself failed to satisfy the McCloy proportionality test.[42] Alternatively, she contends (C) that the ‘intractably broad’ s 13(11) ‘cannot be justified under the second limb’ of Lange.[43] Banerji argues that the provision fails McCloy because it is not suitable (‘[s]ingling out APS employees in the conduct of their private lives in this way lacks a rational explanation’),[44] necessary and/or adequate in balance.[45] On any of these three bases, Banerji submits that the termination of her employment was invalid and the Tribunal’s decision was correct.[46] Several States have intervened, largely in support of the Commonwealth, while the Australian Human Rights Commission (‘AHRC’) has sought leave to appear in an amicus curiae capacity.

III Does the Act Apply to Anonymous Communications by APS Employees?

The first issue to arise in Comcare v Banerji is whether the Public Service Act applies to anonymous communications by APS employees.[47] Discussion of this interpretive issue can be usefully focused on ss 10 and 13(11). While s 15 authorises the sanction (in the present case, termination) and thereby provides the direct burden on the implied freedom, it is s 13(11) — and s 10 by reference to the APS Values — that provide the nexus with political communication. Indeed, s 13(11) arguably burdens the implied freedom even in the absence of a particular s 15 sanction, in light of its chilling effect.[48]

Banerji argues that s 13(11) simply does not apply to anonymous communication (and thereby her conduct). If this contention were accepted, the constitutional issues in the case would fall away and the termination of her employment would be ultra vires on statutory grounds. ‘[T]here must’, she submits, ‘be a nexus between the conduct and the APS as an institution.’[49] Banerji cites the principle of legality and presumption of valid meaning in support of this ‘narrow construction’, decrying that to do otherwise would be to ‘effect an extraordinary intrusion into freedom of expression’.[50] While this may sound superficially compelling, Banerji’s approach lacks sufficient textual grounding. In the author’s view, Banerji’s submissions seek to read in a nexus requirement that has little basis in the terms or context of s 13(11). While the provision may be capable of some reading down, it does not support the categorical bright line immunity that Banerji contends exists for anonymous communications.

The Commonwealth, on the other hand, submits that s 13(11) is wide enough to regulate anonymous communication. However, it is notable that the Commonwealth — like Banerji — otherwise adopts a limited interpretation of the provision’s scope to present a smaller ‘target’ for Lange/McCloy scrutiny. The Commonwealth suggests that, while the provision applies to Banerji’s anonymous comments, this does not mean that it prohibits APS employees from expressing political opinions at all times.[51] Rather, the Commonwealth submits that the provision provides ‘a set of obligations the content of which is context dependent’, with the bounds of the obligations depending on ‘the seniority of the person within the APS’, ‘the person to whom the communication is made’, ‘when and where the communication is made’ and ‘the manner in which the communication is made’.[52] As such, the Commonwealth submits that ‘the Code is not correctly identified as a prohibition on APS members expressing political opinions’ — instead it ‘is more nuanced’.[53] While the Commonwealth’s position is attractive from a policy perspective, as a way of reconciling the competing interests at stake, it arguably replicates the very flaws the Commonwealth points to in Banerji’s construction. Both parties seek to insert limitations that are ‘uncertain’ in ‘content’ and contain ‘no textual foundation’.[54]

Notwithstanding that the Commonwealth and Banerji both advance narrow constructions of s 13(11) (for different purposes), in the author’s view the High Court should give the provision its ordinary meaning. This position is supported by the submissions of Western Australia intervening, where the State’s Attorney-General argues for a broad construction of s 13(11): ‘the words are emphatic in referring to “all times”’.[55] While the Commonwealth sought to rebut this view in its reply submissions — ‘“[a]t all times” in s 13(11) does not mean “always and under any circumstances”’[56] — such linguistic gymnastics strain the interpretation to breaking point. Section 13(11) was drafted in the manner it was — the explanatory memorandum noted that it was intended to be ‘wider than’ its predecessor,[57] and the particular subsection’s application ‘at all times’ is distinguishable from other subsections that are limited to conduct ‘in connection with APS employment’.[58] Together, these indicate a deliberate legislative choice. The Commonwealth should ‘face up to the constitutional consequences’,[59] rather than be permitted to avoid that reality via a specious interpretation. While reading down to avoid constitutional invalidity may be an orthodox statutory interpretation technique, the Court cannot — as it is being asked to do by the Commonwealth — artificially depart from the provision’s ordinary meaning.[60] Where legislation ‘is perfectly clear and entire, free from any ambiguity or omission’, seeking to secure constitutional validity by strained interpretation is not a ‘permissible’ solution.[61]

If the High Court were to give s 13(11) its ordinary meaning, what bearing might this have on the resolution of the dispute? It would, perhaps, strengthen Banerji’s argument that the provision is incompatible with the implied freedom. If s 13(11) does apply to all communications by APS employees, there is more force in Banerji’s attack on the proposition that the Commonwealth can ‘clean[se] APS employees of political opinions’ or limit their ‘ability to express them in ways that do not have a bearing upon the APS as an institution’.[62] It is understandable, then, that the Commonwealth seeks to avoid the natural reading and reasserts its limited approach: ‘The Commonwealth plainly does not suggest that the Code regulates conduct that is “devoid of any connection whatsoever to employment” ... The Respondent erects and demolishes arguments of straw.’[63] On the other hand, if the Commonwealth’s preferred construction were adopted, it would improve the prospects of s 13(11) withstanding Lange/McCloy scrutiny. If the APS Code of Conduct already demands a factorial analysis, with consideration given to the identity of the speaker, the content of the communication and the wider context, concerns about silencing the public service are somewhat tempered.

It may be, and indeed it seems plausible, that the High Court finds s 13(11) does apply to Banerji’s anonymous social media posts, but nevertheless rejects the Commonwealth’s interpretation as too narrow and instead adopts a middle ground. This might favour a finding that the provision is neither valid nor invalid on its face, instead bringing to the fore a ‘difficult category’ where statutory power is apt to be exercised in a manner compatible with the implied freedom.[64] This category raises some complex methodological questions at the intersection of constitutional and administrative law, which remain largely unanswered by existing High Court jurisprudence.

IV Review Methodology

If s 13(11) does apply to Banerji’s anonymous social media posts, the ‘central issue’ in Comcare v Banerji then becomes the review methodology ‘to be applied when a discretionary administrative decision is said impermissibly to burden the implied freedom of political communication’.[65] This, in turn, will be influential in determining the validity of the termination decision. The first methodological issue arising in the submissions is whether the constitutional review task is undertaken entirely at the level of the relevant legislation. The Commonwealth advocates for this approach, with the support of several State Attorneys-General intervening. The AHRC submissions highlight an alternative, that constitutional review could be undertaken directly at the level of the individual decision itself. While the Commonwealth’s approach will typically suffice if the authorising statute is valid or invalid on its face, it is not as straightforward in the difficult category — where the authorising statute confers a power that may be exercised in a manner compatible with the implied freedom. This gives rise to a second methodological issue. In such cases, the Commonwealth argues that the authorising statute must be construed as constitutionally-compliant, and the validity of an individual decision thereunder is resolved as a question of statutory power via administrative law. While determining the boundaries of statutory power may require reference to constitutional limitations, this does not entail undertaking individual-level constitutional review. The third methodological issue raised by the submissions relates to the standard applied when considering the implied freedom in the individual decision context, either when undertaking individual-level constitutional review (per the AHRC on the first issue) or when resolving the statutory boundaries in the difficult category (per the Commonwealth on the second issue).

There is little jurisprudential guidance on evaluating the compatibility of an exercise of executive power with the implied freedom. The difficulty is exacerbated by the implied freedom’s oft-repeated nature: that it is not an individual right. While these questions were considered in an analogous context between 2015 and 2017 in Gaynor, a case involving the political expression of an army reservist, the first instance and intermediate appellate judgments in that case failed to provide a cogent framework.[66] Instead, the limited guidance that does exist springs from Brennan J in Miller v TCN Channel Nine Pty Ltd,[67] a case concerned with discretionary decision-making affecting the freedom of interstate trade protected by s 92 of the Constitution. Justice Brennan stated that ‘[w]here a discretion, though granted in general terms, can lawfully be exercised only if certain limits are observed, the grant of the discretionary power is construed as confining the exercise of the discretion within those [constitutional] limits.’[68] This approach was endorsed by French CJ, Gummow, Hayne, Crennan and Bell JJ in Wotton v Queensland, who noted that as a result ‘any complaint respecting the exercise of power thereunder in a given case ... does not raise a constitutional question, as distinct from a question of the exercise of statutory power’.[69]

These methodological issues will now be considered. To summarise, they are:

(a) Does constitutional review of compatibility with the implied freedom take place at the level of the authorising statute only?

(b) If so, how is compatibility assessed in the difficult category of cases where constitutionality is not clear on the statute’s face?

(c) When the individual decision’s compliance with the implied freedom becomes relevant, either because (a) is answered negatively or because it is necessary in resolving (b), how is this analysis undertaken?

A Constitutional Review to Focus Only on the Authorising Statute?

The Commonwealth contends that when an individual administrative decision is challenged on the basis it impermissibly infringes the implied freedom ‘the question is always whether the legislation that purports to confer the power to make the decision (as opposed to the decision itself) is valid’, by reference to the Lange/McCloy test.[70] The intervening State Attorneys-General echo this approach, arguing variously that ‘the constitutional challenge will necessarily be to the validity of the statutory provisions’[71] and that ‘[t]he implied political freedom is concerned with legislative power, not the facts of particular cases.’[72]

Banerji’s submissions do not dwell on the methodological void and the merits of the respective approaches on this aspect.[73] The AHRC, however, provides a persuasive rebuttal in their amicus curiae submissions. It has long been accepted, at least at the Commonwealth level, that the implied freedom acts as a limit on both legislative and executive power.[74] This dual application is demanded by the ‘structural and systemic imperatives which generate the freedom’ — as the AHRC observes, ‘[t]he constitutionally-prescribed systems [of government] are just as apt to be impeded by executive power’.[75] Accordingly, the AHRC submits that the implied freedom inquiry does not end at the statute because the implied freedom ‘also operates directly on the exercise of s 61 executive power’.[76] On this approach, a constitutional challenge to the exercise of a statutory discretion may be resolved at either the statutory level or at the level of the individual exercise.

B The Implied Freedom and Statutory Discretion

Even on the Commonwealth’s approach, there will be some cases where the compatibility of a statutory discretion with the implied freedom is unclear on the legislation’s face and only becomes apparent following an individual exercise of the discretion — the difficult category.[77] According to the Commonwealth, the Miller/Wotton approach requires that, in such cases, the constitutionality of the authorising statute is accepted and the question becomes whether the discretion’s exercise is within statutory bounds — ‘[t]he discretionary power is construed as extending right up to, but not beyond, the limit of constitutional power’.[78] This approach may require a constitutional analysis to articulate the exact bounds of statutory power, but it does not ‘directly raise any question of constitutional law’ — validity is assumed.[79]

A difficulty with the Commonwealth’s approach is that it elevates what is, it admits, no more than a ‘process of construction’ into an absolute rule.[80] There are long-accepted limits to legislative interpretation. As the High Court said recently, ‘[t]he constructional task remains throughout to expound the meaning of the statutory text, not to remedy gaps disclosed in it or repair it.’[81] The Commonwealth’s reading of the Miller/Wotton approach arguably goes beyond this. As the AHRC submits, ‘[i]t involves reading the text of a statute conferring a broadly-framed discretion as if, by implication, it contained the words “unless the particular exercise of discretion would be contrary to the implied freedom”.’[82] This is a marked departure from the words of s 13(11) and ‘attributes to Parliament an ultimate intention that its laws bear a meaning that is not readily apparent to administrators and citizens’.[83] Such an approach exacts a considerable toll on the rule of law: ‘the law is less accessible ... Parliament is less accountable ... and there is a real risk that the statute will be administered according to its ordinary meaning’.[84] That is not to say that a provision would never demonstrate the necessary intent. But unless the High Court’s further explication of the Miller/Wotton rule involves a significant departure from interpretative orthodoxy, the touchstone will remain legislative intent. That is because, as Gageler J said in NAAJA, ‘a court has no warrant for preferring one construction of a statutory provision over another merely to avoid constitutional doubt’.[85]

C Approach to Review of an Individual Administrative Decision

A third methodological issue arises out of the first and second. To the extent that any analysis focuses on an individual exercise of statutory power (whether for the purposes of determining the boundaries of discretionary power via Miller/Wotton per the Commonwealth or as part of a freestanding constitutional review per the AHRC), how is that analysis undertaken? Does it ‘draw down’ the full Lange/McCloy test for determining compatibility with the implied freedom? The Commonwealth submits that, in asking whether an individual exercise of statutory discretion is compatible, only a limited version of Lange/McCloy must be undertaken. According to the Commonwealth, testing for compatibility and suitability is unnecessary because ‘the compatibility of a statute with those requirements cannot vary from decision to decision’.[86] Instead, the Commonwealth focuses on the adequate-in-balance stage of McCloy. ‘[I]f the statute were to authorise burdens on political communication of the nature and extent that arise from a particular administrative decision purportedly made under the statute’, the Commonwealth proposes as the relevant inquiry, ‘would that present as grossly disproportionate to or as otherwise going far beyond what can reasonably be justified in the pursuit of the statutory purpose?’[87] To the Commonwealth, this approach is attractive because it retains the ‘required systemic focus’.[88]

In contrast, Banerji argues that the Lange/McCloy framework applies within the administrative review process, because ‘proportionality as a “class of criteria” has been applied to administrative decision-making, and no more transparent tools of analysis have been fashioned to date’.[89] This is congruent with the AHRC’s intervention and its emphasis on the implied freedom as limiting executive as well as legislative power. The AHRC, on this point, suggest that the correct approach is to ask ‘whether the particular exercise of power is proportionate or sufficiently tailored to a compatible end’.[90]

Banerji also submits, ‘additionally and not solely’,[91] that orthodox administrative law concepts are relevant: ‘[t]he limit will also be exceeded where the decision-maker fails to consider the implied freedom at all.’[92] Banerji rejects the Commonwealth’s position, drawn from A v Independent Commission Against Corruption,[93] that describing a limit on power as a mandatory relevant consideration is conceptually confused. These differing views are not essential to resolving the present dispute, but they are indicative of administrative law’s failure to develop appropriate tools for enforcing constitutional limits. There is force in Banerji’s submission that administrative decision-making which impacts communicative conduct and does not actively consider the implied freedom, instead complying only inadvertently, ‘has little to commend it’.[94] In the absence of another more appropriate mechanism, utilising relevant considerations to restrain executive power seems attractive — as the aphorism goes, if you only have a hammer, it is tempting to treat everything as a nail. But as a conceptual matter, the implied freedom ‘operates as an ultimate limit on power, and an ultimate limit on power cannot sensibly be described as a mandatory consideration’.[95]

D Consideration

Predicting the High Court’s jurisprudence is a fraught exercise, and there is little guidance as to which methodological approach the bench will take. At the time of writing, it also remains unclear whether the High Court’s reserved judgment in the ‘safe access zone’ cases could change or clarify the implied freedom.[96] While undertaking constitutional review at the level of the authorising legislation has the advantage of familiarity, it is difficult to balance the broadly justifiable aims of s 13(11) of the Public Service Act and the many categories of its reasonable application with the challenging facts of Comcare v Banerji (termination of employment for out-of-hours anonymous political comment). Unless the provision can be read down, constitutional review at the legislative level necessitates an all-or-nothing outcome: s 13(11) is either valid in its entirety (even in extreme cases), or it is invalid despite its many legitimate applications involving no burden on political communication (such as a public servant convicted of child sex offences).[97]

On the other hand, undertaking the constitutional analysis in the context of the individual decision is not free from difficulty either. The wholesale incorporation of Lange/McCloy at an applied level could raise objections that doing so crosses the individual right Rubicon,[98] notwithstanding Banerji’s retort that ‘an individual decision is representative of a larger pattern that would emerge unless the constitutional limitation on the power is enforced’.[99] Although there are no logical inconsistencies inherent in undertaking individual-level review while maintaining the required right/freedom distinction, the High Court’s caution towards any blurring of the distinction may impede the adoption of this approach. The ambiguity of the Court’s comments in Wotton, where they simply recited several of the Commonwealth’s submissions and accepted them, compounds this lack of certainty. In the present case, the Commonwealth criticised the AHRC’s submissions — the AHRC had highlighted the implied freedom’s standalone operation on executive action — as seeking to ‘radically’ limit Wotton.[100] But it is unclear whether the Commonwealth’s expansive reading of Wotton can be sustained.

Alternatively, the High Court could develop a new methodology — this author has elsewhere commended the merits of an ‘as-applied’ constitutional review approach, with origins in Gageler J’s judgment in Tajjour v New South Wales[101] and American jurisprudence.[102] This would see the constitutional analysis take place at the legislative level, but focused on the statutory burden as applied to political communication rather than its general operation. Thus in Tajjour, in the context of Lange’s second limb, Gageler J noted: ‘What is important for that further analysis is that the effective burden ... is confined to the application of the section to an association for a purpose of engaging in [political] communication’.[103] The consequence in Tajjour was that, on Gageler J’s approach, the impugned law was invalid but only in its application to political communication-related activities. This would enable the High Court to invalidate s 13(11) to the extent it impermissibly restricts political communication, without preventing, for example, its continued application in cases of public servants convicted of child sex offences. Whatever the Court’s preference, their choice will have a significant influence on the development of the implied freedom, and likely continue the convergence of constitutional and administrative law. Finally, it is possible, although in the author’s view unlikely, that the Court may sidestep the methodological question, as the Tribunal did. This could be done by resolving the dispute at the interpretative level (as Banerji advocates), or by holding that the termination is valid or invalid on any approach and therefore the correct methodology need not be determined.

V Validity of the Termination Decision

Unless the High Court accepts Banerji’s contention that her social media activity does not fall within the scope of s 13(11), the bench will — irrespective of the methodology adopted — be required to determine the validity of the termination of her employment. The chosen methodology may have a significant impact on the outcome of that evaluative exercise, but the Court must ultimately address whether terminating the employment of a public servant for expressing political opinion anonymously on Twitter is compatible with the Constitution by reference to Lange/McCloy. Reasonable minds can and likely will differ on this issue.

In the author’s view, the Court should not uphold broad intrusions into the private lives of public servants, a consequence of which is the distortion of political discourse among a significant portion of the Australian polity. The salient facts of Banerji are worth restating. Banerji was a mid-level public servant (APS6), terminated because of the content and tone of her political communications, which were made anonymously, in her own time[104] and using her own electronic devices. That Banerji’s identity was subsequently discovered does not change the complexion of her prior tweets, particularly where, as the Tribunal noted, ‘it was the Department itself which dissolved her anonymity’.[105]

It is possible to imagine alterative versions of this fact pattern where the validity analysis is more likely to be answered in the Commonwealth’s favour: if the tweeter was a senior public servant, such that the government might lose trust in the relevant department, or if the tweets were sent during work hours. Likewise, it possible to envisage scenarios where validity is even more questionable: if the tweets concerned policy relating to a different department, and contained mild analysis rather than inflammatory criticism.[106] Between these poles, Banerji sits closer to the latter end of the spectrum. Prohibiting anonymous political expression by non-senior public servants on a topic of immense national interest (as the ongoing debate about border protection ahead of the 2019 Federal Election demonstrates) is not adequate in the balance it strikes between the competing policy objectives. Instead, as the Tribunal commented, ‘restrictions in such circumstances bear a discomforting resemblance to George Orwell’s thoughtcrime’.[107]

There is also a concerning trend in the handful of litigated cases arising in the present or analogous contexts over the past decade or two: all involved criticism of Commonwealth policy. But the obligation to act in an apolitical manner cuts both ways. As Western Australia highlights in its intervening submissions, s 13(11) ‘imposed a substantive limit on the ability of the employee to promote or criticise’ government policy.[108] It is notable that the Commonwealth has made no publicised attempts to discipline public servants for publicly promoting government policy. Indeed, the Australian Public Service Commission’s latest guidance, Making Public Comment on Social Media: A Guide for Employees, advises that while ‘[c]riticising the work, or the administration, of your agency is almost always going to be seen as a breach’ of the APS Code of Conduct, it ‘doesn’t stop you making a positive comment on social media about your agency’.[109] This somewhat undermines the Commonwealth’s submissions emphasising the upmost importance of public sector impartiality, as a legitimate aim for the purposes of Lange. It might be suggested that had Banerji instead praised Commonwealth immigration policy (anonymously or otherwise), she would not have found herself in this current predicament.

While the implied freedom might be uniquely Australian, the High Court is not alone in being asked how to appropriately balance the personal expressive rights of public servants (and the public interest in the dissemination of their political views) with the need for an impartial and apolitical public service. It is a vexing dilemma that has confronted superior courts in Europe,[110] the United States,[111] and Canada,[112] among other jurisdictions. Although the exhortation that the implied freedom is not a personal right has greatly restricted reliance on comparative jurisprudence, international case law may provide helpful guidance in Comcare v Banerji — or at the very least ‘food for thought’.[113] The High Court might, for example, wish to heed the Canadian Supreme Court’s warning: whatever the exact balance struck, public servants ‘cannot be ... “silent members of society”’.[114]

VI Conclusion

It has long been said that hard cases make bad law.[115] Comcare v Banerji is a hard case. Its resolution requires the High Court to determine whether the legislature has struck a constitutionally-compatible balance between conflicting — and compelling — policy considerations. In reaching its conclusion, the Court must explicate the appropriate test for assessing the compliance of administrative decisions with the implied freedom of political communication. On neither front is the Court helped by an abundance of domestic cases, while its insistence on the unique nature of the implied freedom inhibits reliance on comparative law. Whatever the outcome, Comcare v Banerji will leave a significant legacy — for tweeting public servants

(at both federal and state level) and for the broader development of implied freedom jurisprudence.


[*] Visiting Fellow, Centre for International and Public Law, ANU College of Law, Canberra; Legal Advisor, Legal Policy & Research Unit, International Bar Association, London, England. The author acknowledges the helpful comments of: Professor James Stellios and Associate Professor Heather Roberts, both of ANU College of Law; Emily Hammond, University of Sydney Law School; and an anonymous reviewer.

[1] Edward Hughes, ‘Sir James Stephen and the Anonymity of the Civil Servant’ (1958) 36 Public Administration 29, 29–36.

[2] R Plehwe, ‘Political Rights of Victorian Public Employees’ (1983) 42(3) Australian Journal of Public Administration 362, 365–6.

[3] Commonwealth Public Service Regulations 1902 (Cth) reg 41.

[4] Bennett v President, Human Rights and Equal Opportunity Commission [2003] FCA 1433; (2003) 134 FCR 334 (‘Bennett’). The validity of the replacement restriction was upheld in R v Goreng Goreng [2008] ACTSC 74; (2008) 220 FLR 21. See generally Nicola Bevitt, ‘The Australian Border Force Act 2015 (Cth) Secrecy Provisions — Borderline Unconstitutional’ [2017] SydLawRw 12; (2017) 39(2) Sydney Law Review 257.

[5] de Freitas v Permanent Secretary of Ministry of Agriculture, Fisheries, Lands and Housing [1998] UKPC 30; [1999] 1 AC 69, 75–6 (‘de Freitas’).

[6] Attorney-General (Cth), ‘Submissions of the Attorney-General of the Commonwealth (Intervening)’ in Comcare v Banerji, Case No C12/2018, 7 November 2018, [26] (‘Attorney-General (Cth) Submissions’), quoting Australian Public Service Commission, APS Values and Code of Conduct in Practice (August 2017) 8.

[7] R S Parker, ‘Official Neutrality and the Right of Public Comment: II. The Vow of Silence’ (1964) 23(3) Australian Journal of Public Administration 193, 196.

[8] The contractual element has not received sufficient scholarly attention, partly because the statutory overlay directs primary attention to the constitutional question. All employment contracts contain an implied duty of fidelity that the employee will act in the employer’s best interests: Robb v Green [1895] UKLawRpKQB 131; [1895] 2 QB 315. The extent to which this duty might adapt to the implied freedom was highlighted, but not determined, in Bennett: [2003] FCA 1433; (2003) 134 FCR 334, 362–4.

[9] In June 2018 there were 1 987 000 public sector employees in Australia, constituting approximately 16 per cent of the Australian workforce: Australian Bureau of Statistics, Employment and Earnings, Public Sector, Australia, 2017–18 (8 November 2018) <http://www.abs.gov.au/ausstats/abs@.nsf/

mf/6248.0.55.002>; Australian Bureau of Statistics, Labour Force, Australia, Jan 2019 (21 February 2019) <http://www.abs.gov.au/ausstats/abs@.nsf/mf/6202.0> .

[10] San Diego v Roe, [2004] USSC 5997; 543 US 77, 80 (2004).

[11] Australian Human Rights Commission (‘AHRC’), ‘Submissions of the AHRC Seeking Leave to Appear as Amicus Curiae’, Submission in Comcare v Banerji, Case No C12/2018, 12 December 2018, [24] (‘AHRC Submissions’), quoting Commonwealth Public Service Board, Guidelines on Official Conduct of Commonwealth Public Servants (1979) 47.

[12] Paul Finn, Abuse of Official Trust: Conflict of Interest and Related Matters, Integrity in Government Second Report (Australian National University, 1993) 61.

[13] See generally, Comcare, ‘Appellant’s Chronology’, Submission in Comcare v Banerji, Case No C12/2018, 7 November 2018 (‘Appellant’s Chronology’).

[14] Markus Mannheim, ‘Public Servant Loses Fight over Twitter Attack on Government’, Canberra Times (online), 13 August 2013 <https://www.canberratimes.com.au/national/public-servant-loses-fight-over-twitter-attack-on-government-20130812-2rsgn.html?_ga=2.106608707.1931213015.154

1879927-367294655.1447915721>.

[15] The Tribunal found that, of over 9000 tweets, one had been made during work hours. Yet the Tribunal held that Banerji had ‘been careful, even assiduous, in avoiding posting tweets during working hours’ and that ‘nothing turns’ on the finding regarding the lone tweet: Re Banerji and Comcare (Compensation) [2018] AATA 892 (16 April 2018) [26], [30].

[16] A cognate obligation is now found in Public Service Act s 10(5), following minor legislative reform.

[17] Banerji v Bowles [2013] FCCA 1052 (9 August 2013) (‘Banerji’).

[18] (2001) 208 CLR 199, 280–82.

[19] Banerji [2013] FCCA 1052 (9 August 2013) [25].

[20] Ibid [101].

[21] Ibid [104].

[22] Appellant’s Chronology, above n 13, 5.

[23] [1997] HCA 25; (1997) 189 CLR 520 (‘Lange’). See generally a helpful discussion of the implied freedom in Shireen Morris and Adrienne Stone, ‘Before the High Court – Abortion Protests and the Limits of Political Communication: Clubb v Edwards; Preston v Avery[2018] SydLawRw 15; (2018) 40(3) Sydney Law Review 395.

[24] [2004] HCA 39; (2004) 220 CLR 1 (‘Coleman’).

[25] [2015] HCA 34; (2015) 257 CLR 178 (‘McCloy’).

[26] [2017] HCA 43; (2017) 261 CLR 328 (‘Brown’).

[27] This formulation merges passages from McCloy and Brown: McCloy [2015] HCA 34; (2015) 257 CLR 178, 195–6 (French CJ, Kiefel, Bell and Keane JJ); Brown [2017] HCA 43; (2017) 261 CLR 328, 364. In January 2019 (after the filing of submissions in the present dispute), the High Court decided Unions NSW v New South Wales [2019] HCA 1 (29 January 2019). While the various judgments do not significantly alter the implied freedom orthodoxy, they do collectively stress the importance of the Court being satisfied that the burden on the implied freedom is necessary to achieve the law’s legitimate purpose.

[28] Re Banerji and Comcare (Compensation) [2018] AATA 892 (16 April 2018).

[29] Ibid [71], [74].

[30] Ibid [113].

[31] Ibid [115].

[32] Ibid [117].

[33] Ibid.

[34] Ibid [128].

[35] Ibid.

[36] Ibid [126].

[37] Ibid [125] (emphasis in original).

[38] Ibid [127].

[39] Attorney-General (Cth) Submissions, above n 6, [11].

[40] Ibid [8]. There is a second limb to this argument: to the extent that the validity is dependent on the severity of sanctions imposed, s 13(11) remains valid in all its applications because the statute properly construed requires sanctions be proportionate, which ‘is sufficient to ensure that the scheme as a whole remains within constitutional limits’: ibid [9]. Due to brevity requirements, this aspect will not be considered further.

[41] Ibid [10]–[11].

[42] Michaela Banerji, ‘Respondent’s Submissions’, Submission in Comcare v Banerji, Case No C12/2018, 5 December 2018, [43]–[50] (‘Banerji Submissions’).

[43] Ibid [3].

[44] Ibid [58].

[45] Ibid [59]–[60].

[46] Ibid [17], [61].

[47] This is raised on a notice of contention filed by Banerji, and the Commonwealth has submitted that the Court should not deal with it on procedural grounds: Attorney-General (Cth), ‘Reply Submissions of the Attorney-General of the Commonwealth’ in Comcare v Banerji, Case No C12/2018,

19 December 2018, [2] (‘Attorney-General (Cth) Reply Submissions’).

[48] See generally Frederick Schauer, ‘Fear, Risk and the First Amendment: Unraveling the “Chilling Effect”’ (1978) 58 Boston University Law Review 685.

[49] Banerji Submissions, above n 42, [17].

[50] Ibid [39]–[41].

[51] Ibid.

[52] Attorney-General (Cth) Submissions, above n 6, [22].

[53] Ibid.

[54] Attorney-General (Cth) Reply Submissions, above n 47, [3].

[55] Attorney-General (WA), ‘Annotated Written Submissions on Behalf of the Attorney General for Western Australia (Intervening)’ in Comcare v Banerji, Case No C12/2018, 14 November 2018, [12] (‘Attorney-General (WA) Submissions’).

[56] Attorney-General (Cth) Reply Submissions, above n 47, [3].

[57] Explanatory Memorandum, Public Service Bill 1999 (Cth) 25.

[58] Public Service Act ss 13(1)–(4).

[59] North Australian Aboriginal Justice Agency Ltd v Northern Territory [2015] HCA 41; (2015) 256 CLR 569, 610 (Gageler J) (‘NAAJA’).

[60] International Financial Trust Co Ltd v NSW Crime Commissioner [2009] HCA 49; (2009) 240 CLR 319, 349 (French CJ).

[61] de Freitas [1998] UKPC 30; [1999] 1 AC 69, 77.

[62] Banerji Submissions, above n 42, [55].

[63] Attorney-General (Cth) Reply Submissions, above n 47, [3] n 3.

[64] James Stellios, ‘Marbury v Madison: Constitutional Limitations and Statutory Discretions’ (2016) 42(3) Australian Bar Review 324, 331

[65] Attorney-General (Cth) Submissions, above n 6, [4].

[66] Gaynor v Chief of the Defence Force (No 3) [2015] FCA 1370; (2015) 237 FCR 188; Chief of Defence Force v Gaynor (2017) 246 FCR 298 (‘Gaynor’). Special leave to appeal was ultimately refused by the High Court: Transcript of Proceedings, Gaynor v Chief of the Defence Force [2017] HCATrans 162 (18 August 2017). See generally Joshua Forrester, Lorraine Finlay and Augusto Zimmermann, ‘Finding the Streams’ True Sources: The Implied Freedom of Political Communication and Executive Power’ (2018) 43 University of Western Australia Law Review 188.

[67] [1986] HCA 60; (1986) 161 CLR 556 (‘Miller’).

[68] Ibid 614, quoting Inglis v Moore (No 2) (1979) 46 FLR 470, 476 (St John and Brennan JJ).

[69] (2012) 246 CLR 1, 14 (‘Wotton’).

[70] Attorney-General (Cth) Submissions, above n 6, [4] (emphasis in original).

[71] Attorney-General (NSW), ‘Submissions of the Attorney General for New South Wales (Intervening)’, Submission in Comcare v Banerji, Case No C12/2018, 14 November 2018, [18].

[72] Attorney-General (WA) Submissions, above n 55, [23].

[73] Banerji Submissions, above n 42, [42]–[51].

[74] See, eg, Lange [1997] HCA 25; (1997) 189 CLR 520, 560.

[75] AHRC Submissions, above n 11, [14].

[76] Ibid [17] (emphasis in original).

[77] Stellios, above n 64, 331.

[78] Attorney-General (Cth) Submissions, above n 6, [50].

[79] Ibid (emphasis in original).

[80] Ibid.

[81] HFM043 v Republic of Nauru [2018] HCA 37; (2018) 92 ALJR 817 [24] (Kiefel CJ, Gageler and Nettle JJ).

[82] AHRC Submissions, above n 11, [47].

[83] Ibid [49].

[84] Ibid [45].

[85] [2015] HCA 41; (2015) 256 CLR 569, 604.

[86] Attorney-General (Cth) Submissions, above n 6, [52].

[87] Ibid [56].

[88] Ibid.

[89] Banerji Submissions, above n 42, [43].

[90] AHRC Submissions, above n 11, [58].

[91] Banerji Submissions, above n 42, [45] (emphasis altered).

[92] Ibid.

[93] [2014] NSWCA 414; (2014) 88 NSWLR 240, 257 (Basten JA).

[94] Banerji Submissions, above n 42, [45].

[95] Bret Walker and David Hume, ‘Broadly Framed Powers and the Constitution’ in Neil Williams (ed), Key Issues in Public Law (Federation Press, 2018) 144, 157.

[96] See Transcript of Proceedings, Clubb v Edwards; Preston v Avery [2018] HCATrans 210 (11 October 2018). See also Morris and Stone, above n 23.

[97] Cooper v Australian Taxation Office [2014] FWC 7551 (6 November 2014); Cooper v Australian Taxation Office [2015] FWCFB 868 (11 March 2015).

[98] The High Court’s insistence that the implied freedom is not a personal right, with the implications that follow, has been criticised. See, eg, Adrienne Stone, ‘Rights, Personal Rights and Freedoms: The Nature of the Freedom of Political Communication’ [2001] MelbULawRw 13; (2001) 25(2) Melbourne University Law Review 374.

[99] Banerji Submissions, above n 42, [44].

[100] Attorney-General (Cth) Reply Submissions, above n 47, [14] n 10.

[101] [2014] HCA 35; (2014) 254 CLR 508 (‘Tajjour’).

[102] Kieran Pender, ‘“Silent Members of Society”? Public Servants and the Freedom of Political Communication in Australia’ (2018) 29(4) Public Law Review 327, 345–7. See generally Richard H Fallon Jr, ‘As-Applied and Facial Challenges and Third-Party Standing’ (2000) 113 Harvard Law Review 1321.

[103] [2014] HCA 35; (2014) 254 CLR 508, 583.

[104] See above n 15.

[105] Ibid [93].

[106] It must be remembered, though, that the implied freedom ‘does not protect only the whispered civilities of intellectual discourse’, but extends to ‘insult and emotion, calumny and invective’ Coleman [2004] HCA 39; (2004) 220 CLR 1, 91 [239] (Kirby J).

[107] Re Banerji and Comcare (Compensation) [2018] AATA 892 (16 April 2018) [116] (emphasis in original).

[108] Attorney-General (WA) Submissions, above n 55, [13] (emphasis added).

[109] Australian Public Service Commission, Making Public Comment on Social Media: A Guide for Employees (2017) 3.

[110] Wille v Liechtenstein [1999] ECHR 107 (28 October 1999); Haseldine v United Kingdom (1992) D&R 73; Ahmed v United Kingdom [1998] ECHR 78 (2 September 1998); Redfearn v United Kingdom [2012] ECHR 1878 (6 November 2012).

[111] Pickering v Board of Education, [1968] USSC 125; 391 US 563 (1968); San Diego v Roe, [2004] USSC 5997; 543 US 77 (2004); Lane v Franks 134 S Ct 2369 (2014).

[112] Fraser v Public Service Staff Relations Board [1985] 2 SCR 455 (‘Fraser’); Haydon v Canada [2001] 2 FC 82; Haydon v Canada [2004] FC 749; Osborne v Canada [1991] 2 SCR 69.

[113] Transcript of Proceedings, Barclay v Penberthy [2012] HCATrans 98 (1 May 2012) 2123 (Crennan J).

[114] Fraser [1985] 2 SCR 455, 466.

[115] Northern Securities Company v United States, [1904] USSC 64; 193 US 197, 400 (Holmes J) (1904).