Why is it so hard to take lawful strike action in Australia?

How has the right to strike been enacted?

Under the common law, all strike action (including action short of a strike) is unlawful (see Deakin and Morris, 2009: 899–900; Stewart et al., 2016: ch. 26). An employee who fails to obey a lawful and reasonable command of their employer in support of an industrially motivated agenda will repudiate their contract and be liable to dismissal. This ‘unlawful act’ when committed collectively or when induced by a trade union will form the basis of the commission of an economic tort. The economic torts are a group of potential liabilities arising at common law that apply when an unlawful act (like breach of an employment contract) is carried out in concert with others. The economic torts enable employers to obtain interlocutory injunctive relief against threatened strike action, stopping proposed action under threat of contempt of court for the workers and organisers involved, or, if pursued to judgment, damages for the harms caused (see further McCrystal, 2010b: ch. 5; Stewart et al., 2016: 954–959).

Strike action cannot be lawful unless the impact of the common law has been nullified through statutory enactment. In Australia this has been implemented by granting an immunity to industrial action which complies with the prerequisites. Under s 415, no action lies under any law (written or unwritten) in relation to protected industrial action unless that action involves personal injury or property damage or theft. Employees are protected against adverse action in the form of dismissal or prejudice in employment for participating in protected industrial action under ss 340–342. Where employees and their representatives fail to comply with the statutory rules, the common law rules apply – even where non-compliance is mistaken or inadvertent. The stakes for non-compliance are potentially high. In addition to any relevant statutory sanctions, employees may lose their jobs, and unions can be pursued for the losses experienced by the employer.

However, a more insidious problem is the well-known presumption of statutory interpretation that Parliament would not intend to overthrow fundamental principles, infringe rights, or depart from the general system of law, without ‘expressing its intention with irresistible clearness’ (Potter v Minahan at 304; Coco v The Queen at 437; see further Pearce and Geddes, 2014: 236). Where a statute appears to override common law rights, the court will not presume that this is the intended effect unless this intention is clearly expressed. This presumption of interpretation has been used by the courts to construe provisions of the FW Act in a manner that protects common law rights and remedies over an approach that would protect the right of workers to take lawful industrial action. The most recent example is the 2017 decision of the High Court in Esso Australia v The Australia Workers’ Union (Esso). Faced with two competing interpretations of a provision of the FW Act, the High Court by majority preferred the one that narrowed the scope of protected industrial action, referring to the immunity from civil suit provided by the FW Act as a ‘privilege’ (Esso at 423). Over time, this approach to interpretation of the scope of the protected industrial action provisions has made it cumulatively harder for workers to strike (McCrystal, forthcoming).

A further problem is that the protection of individual employees who take protected industrial action has been seriously denuded by adverse judicial interpretation. In BHP Coal Pty Ltd v CFMEU (BHP Coal), an employee was terminated after his conduct on a picket was considered by the employer to be disrespectful and in breach of the company’s code of conduct. While the employee was not engaged in industrial action at the time because, as discussed in the following, a picket is not industrial action under the FW Act, the High Court found that he was engaged in lawful and peaceful industrial activities in the form of protest action. However, he was not protected from adverse action (which includes dismissal and prejudicial treatment short of dismissal) under the FW Act. ¹ According to the High Court, he was terminated by the employer because he breached the company’s code of conduct, not because of his industrial activities. In the light of BHP Coal, it is plausible to suggest that an employee engaged in protected industrial action would not be protected against adverse treatment under the FW Act if the adverse treatment occurred because of the impact of their absence from work, ² or because the employer found the manner of their engagement in the action to breach company policy irrespective that the mere fact of their participation in industrial action is protected (see Riley, 2018).

What objectives may be supported by lawful strike action?

Under the FW Act lawful strike action by employees may only be taken against an employer in support of claims to be included in a proposed single-enterprise agreement (s 409(1)). Employees and their representatives cannot use industrial action to pursue social or economic objectives that cannot otherwise be quantified as claims in an enterprise agreement. They cannot respond to work health and safety concerns at the workplace with industrial action. ³ They cannot use industrial action to try to resolve a dispute with an employer over compliance with industrial standards. Lawful strike action cannot be taken during the currency of an existing enterprise agreement, even in respect of claims that are not otherwise covered by that agreement (s 417). Lawful strike action cannot be taken in pursuit of ‘unlawful terms’ (s 409(3)), which include inter alia ‘discriminatory’ terms, claims for bargaining services fees, or attempts to give employees earlier access to unfair dismissal protection (s 194). Lawful strike action cannot be taken in support of a multi-enterprise agreement, a greenfields agreement, in support of common claims being made by workers in the same union engaged at different enterprises (pattern bargaining) (s 421), or in sympathy with other workers at a different enterprise. Secondary boycotts are prohibited under the Competition and Consumer Act 2010 (Cth) and potentially subject to large fines. ⁴

The capacity to take lawful industrial action is limited to the window of opportunity that applies after an existing agreement has expired and before a new one has been made. During that time, protected industrial action may be taken but only at that enterprise, against that employer, and only to support claims that can potentially be included in an enterprise agreement under the FW Act. This has meant that the scope of claims capable of inclusion in an enterprise agreement has been the subject of ongoing political interest – with the narrowest list of claims, and the longest list of ‘prohibited content’ applying during the Work Choices period from 2006 to 2009 after amendments passed by the Workplace Relations (Work Choices) Amendment Act 2005 (Cth) (see Stewart and Riley, 2007).

While the current scope of permissible claims is broader than during the Work Choices era, the legislative formula of ‘matters pertaining to the employment relationship’ used in s 172(1)(a) to identify content capable of inclusion in an enterprise agreement is the subject of highly confusing and inconsistent case law (see Harris, 2006; Stewart et al., 2016: 365–373). If challenged, employees and their bargaining representatives must show that the claims supported by industrial action are capable of inclusion in an agreement or that they ‘reasonably believe’ the claims to be capable of inclusion. This test has provided employers with the opportunity to intervene to challenge and delay lawful industrial action by employees and their bargaining representatives (see the discussion of the Australia Post cases in McCrystal, 2012: 146–149), and the ambiguities around the statutory definition continue to catch some bargaining representatives out (as in United Voice v Castlemaine Perkins Pty Limited T/A Lion). Given that this is the main determinant of the claims that may be supported by lawful industrial action, the failure of successive Parliaments to legislate for the content of enterprise agreements in a coherent manner continues to be a source of difficulty.

Finally, the absolute prohibition in the FW Act on strike pay should be noted. Under the common law, the principle of ‘no work, no pay’ reflects the understanding of a contract of employment as one under which the service of an employee earns wages, provided that the employee is ready, willing and able to perform the entirety of their contractual obligations (Automatic Fire Sprinklers v Watson). If services are not offered in full, the obligation to pay wages does not arise. However, nothing under the common law prevents the creation of a subsequent agreement to pay lost wages as part of the settlement of a dispute. By contrast, the FW Act expressly prohibits any such settlement and makes unlawful any attempt to pay or receive wages for a period of industrial action (ss 470(1), 473, 474, 475). Different rules apply in respect of partial work bans where an employer may choose to pay in full, pay a proportion reflecting the value of the work provided, or not to pay at all (s 471). However, as with all industrial action, payment for partial work bans cannot be included as part of the settlement of a dispute – the choice to pay for partial bans is at the discretion of the employer.

This further restricts the scope of permissible claims that may be sought in an enterprise agreement and impacts the economic capacity of unions to fund industrial action by including lost wages claims as part of industrial settlements.

What are the prerequisites for lawful industrial action?

When it is lawful for employees and their representatives to take industrial action, the legislative immunity only attaches if the statutory prerequisites are complied with. Most are procedural, but some are substantive and may present insurmountable obstacles.

Before protected industrial action can be taken, the notification time must have passed (ss 173(2), 437(2A)). This means that bargaining for a new enterprise agreement must have commenced, either because the employer has agreed to bargain or because the employer has been required to bargain by an order of the FWC, most commonly a majority support determination (s 172). Before 2015, protected industrial action could be taken by employees and their representatives in support of a proposed agreement where bargaining had not commenced (JJ Richards & Sons Pty Ltd v Fair Work Australia; discussed in Howard, 2013). However, this was amended in 2015 by the Fair Work Amendment Act (Cth) such that employees and their representatives can no longer seek to influence the commencement of bargaining through industrial action. This has further restricted the circumstances in which employees can lawfully take industrial action (see further Creighton, 2018).

If the notification time has passed, the bargaining representative organising and/or engaging in the industrial action must not have contravened any orders that apply to them in relation to that bargaining round (s 413(5)). It had previously been thought that this meant that the bargaining representative could not be in breach of an order at the time they took any action. However, the High Court in Esso found that the bargaining representatives cannot have breached any order of the FWC or a court – however inadvertent, mistaken or technical the breach – and remain eligible to take protected industrial action in relation to bargaining for that agreement. According to the High Court, this ‘is to ensure that persons who have shown they cannot be trusted to comply with orders … are not to be trusted with the immunity afforded in relation to protected industrial action’ (Esso at 422–423).

The effect of the High Court decision in Esso is to create an industrial environment where bargaining representatives must pay scrupulous attention to any orders made in connection with bargaining (good faith orders, service orders), or access to protected industrial action may be lost for the entirety of the bargaining round. Justice Gageler in dissent noted that the FW Act scheme of bargaining and industrial action allows for such a large number of orders that breaches must be expected, and often may be easily remedied (Esso at 430). Consequently, he argued that the majority’s approach created a ‘harsh and rigid form of industrial discipline’ likely to create industrial cripples and outlaws (Esso at 435). Justice Gageler found that this was not consonant with a statutory scheme designed to create an environment for collective bargaining that is ‘fair, flexible and efficient’ (Esso at 435).

To fall within the scope of protected industrial action, any action taken must be ‘industrial action’ within the definition of that term in s 19(1). The scope of this section is intentionally broad, picking up a wide range of industrially motivated conduct including total work stoppages, work bans and the performance of work in a manner different from that in which it is normally performed. However, it has been held that the definition does not include a picket undertaken by employees while otherwise engaged in protected industrial action (David’s Distribution v NUW), which means that the picket itself and any conduct thereupon does not constitute industrial action and may therefore be subject to injunctive relief or further common law action. It is also likely that the more creative forms of industrial action undertaken by unions in recent years, such as sending emails with the CAPS LOCK function on, wearing union T-shirts, or taking lunch breaks in management offices, do not constitute industrial action (see e.g. United Firefighters Union of Australia v Easy; Ambulance Victoria v United Voice). The definition in s 19(1) requires the relevant action to have the effect of restricting, limiting or delaying the performance of work, something that such actions may not actually do in practice (Creighton et al., 2017: 396–402). This narrows the scope of action available to unions, especially when they are trying to minimise the overall cost of taking industrial action.

The next requirement is that unless action is taken in response to employer industrial action, protected industrial action must be authorised by the relevant employees by a protected action ballot (s 409(2)). To conduct a ballot, a bargaining representative must first apply to the FWC for a ‘protected action ballot order’ (a PABO), which gives permission for a ballot to be conducted (the history of this provision is outlined in Creighton et al., 2016).

The PABO provisions are contained in Part 3-3, Division 8 of the FW Act. A bargaining representative must apply to the FWC for an order (s 437) and the FWC must make an order if it is satisfied that the application is made in accordance with the Act, the notification time has passed, and the applicant has been, and is, genuinely trying to reach an agreement at the relevant enterprise (s 443). Employers have the right to oppose a PABO application, and frequently make use of this right (Australia Postal Corp v CEPU; Creighton et al., 2018a: 64–66). This can lead to considerable delay in obtaining a PABO.

Once a PABO is ordered, a ballot must be conducted by a protected action ballot agent who is independent of the bargaining representative. If the agent is the Australian Electoral Commission (AEC) then the cost of the ballot is borne by the Commonwealth, but if the bargaining representative uses someone else, they pay the full cost (ss 464(2), 465(2)). The electoral roll consists of those who are represented by the bargaining representative and will be covered by the proposed agreement (s 453). The ballot must be secret and may be conducted by postal ballot, attendance ballot or electronic ballot (s 450(2)). In order for proposed industrial action to be approved, a majority of those on the electoral roll must cast a vote and a majority of those who vote must approve the proposed action (s 459(1)).

The PABO and ballot process contributes significantly to making it harder for employees to take lawful industrial action. First the PABO process increases delay, provides an opportunity for employers to intervene, and requires bargaining representatives to establish that they are complying with a statutory test that is not routinely applied when employers take industrial action. Second, the ballot itself creates problems. The evidence suggests that while employees do not generally vote against proposed industrial action, they do fail to vote. Achieving quorum is the most common reason that proposed action is not approved in a protected action ballot (Creighton et al., 2018a: 67). In the vast majority of non-quorate ballots, the proposed action would have been approved but for the quorum requirement (Creighton et al., 2018b). Another complication in respect of the PABO process is the form of ballot used. The AEC does not offer electronic balloting, and most commonly offers postal ballots over attendance ballots (Creighton et al., 2018b). Postal ballots fail to meet quorum at a rate over three times as often as attendance ballots (Creighton et al., 2018b). The cumulative effect is that the administrative and economic costs of the PABO process drive bargaining representatives to use the AEC and this most commonly leads to the use of a postal ballot, the method that is least efficacious in terms of reaching quorum.

The next prerequisite to lawful industrial action is notice. Bargaining representatives must give 3 working days’ notice of proposed industrial action (not including the day the notice is given or the day the action commences, which means that it is really 5 days) or a longer period (up to 7 days) if the FWC requires it (s 414). This provides a further opportunity for employers to intervene and delay the process by seeking a longer period of notice, which occurs in around 27.5% of PABO applications (Creighton et al., 2018a: 66), or by challenging the notice itself on the basis inter alia that the industrial action specified is too vague or does not actually constitute industrial action.

Further, protected industrial action must be commenced within 30 days (or 60 if extended by the FWC) of the declaration of the ballot results (FW Act s 459(1)). This section has been interpreted to mean that if the ballot approves multiple forms of action, the protection will only extend after the initial 30 days to those actions taken in that period. The ability to take any form of action not used is lost (United Collieries Ltd v CFMEU; Energy Australia Yallourn Pty Ltd v CFMEU). Further, if the union wants to use any form of industrial action again, the ballot must have approved multiple occurrences of that action (see further Denvir and McCrystal, 2017). This introduces a level of technical artifice into the provisions, forcing employees to rush to take one instance of all approved action in the first 30 days, or else lose the ability to use it and have to re-ballot with attendant cost and delay.

Finally, protected industrial action can only proceed if the action that is proposed to be taken is the action notified to the employer and authorised in the ballot. In Esso, the primary dispute between the employer and the Australian Workers' Union (AWU) was a factual one over exactly what was entailed by the notice of industrial action given to the employer. The AWU contended that in giving notice of ‘deisolation of equipment’ this included all the steps necessary to deisolate equipment. The employer claimed that it only involved the last step of turning the equipment back on. During the litigation, two judges found in favour of the union’s assertion, two found in favour of the employer – but the employer’s interpretation prevailed, with the result that the industrial action was unprotected (the history of the litigation is outlined in Creighton and McCrystal (2017)).

The difficulty with the outcome in Esso is that the union genuinely believed it was taking protected industrial action, took all the right steps and had a reasonable basis to conclude that the action undertaken was the action it had approved in the ballot and had provided notice of to the employer. Here, legal protection for the union and the workers concerned was lost because the majority of the Full Federal Court disagreed on the facts of the case. This is not a rational way for critical protections against common law liability to operate and leaves unions unable to be satisfied with certainty that the action they are taking is protected.

When can lawful strike action be stopped?

The next reason it is so hard for Australian employees to take industrial action when they need to is the relative ease with which protected industrial action can be stopped. Where bargaining representatives and employees have complied with the statutory prerequisites they are free to take protected industrial action, which theoretically can continue unabated until agreement is reached.

However, the FWC has the power to bring protected industrial action to an end on four grounds. It may suspend industrial action where one of the parties requests a cooling-off period (s 425) or where industrial action is causing ‘significant harm to a third party’ (s 426). It may suspend or terminate industrial action (with termination leading to arbitration of the dispute) where significant harm is being caused to the bargaining parties themselves (s 423) or where the industrial action threatens, or would threaten, ‘to endanger the life, the personal safety or health, or the welfare of the population or part of it’ or to ‘cause significant economic damage to the Australian economy or an important part of it’ (s 424).

The first three grounds have not caused significant concern in respect of the ability to take protected industrial action (McCrystal, 2018). However, s 424 is a different story. In the Qantas dispute, the employer engineered a lockout of employees to trigger termination of their own industrial action under s 424. The effect of this was to end the right of the employees concerned to take protected industrial action because suspension or termination of any industrial action during bargaining removes the right of other parties to bargaining to access protected industrial action (s 413(7)(a)). In the Qantas example, the unions concerned had taken great care with their own industrial action to avoid triggering a suspension or termination under s 424, but they could not stop the employer from responding with their own industrial action which deliberately had this effect (Forsyth and Stewart, 2013).

Of even greater concern is the requirement in s 424 that the FWC suspend or terminate protected industrial action that ‘threatens to endanger the welfare’ of the population or part of it. The section has been used against Monash University academics in respect of bans on releasing student results. Finding that the result bans threatened the welfare of university students, even though there was an exemption process in place for students who required their results, the FWC suspended the industrial action for long enough for all results to be released, undermining the effect of the action (NTEU v Monash University). This same ground was used to suspend a threatened strike by Sydney Trains employees in February 2018, on the basis of the threat the action posed to the welfare of Sydney commuters (Sydney Trains; NSW Trains; The Hon. Dominic Perrottet, Minister for Industrial Relations (NSW)). Given that almost all successful industrial action in any services sector has an impact on third parties, this ground of suspension has the potential to bring to an end a whole range of lawful industrial action, particularly when that action does as intended and causes disruption (McCrystal, 2018).

What happens if there is failure to meet the requirements for lawful strike action?

The final component of this regulatory picture is the consequence of failing to take lawful industrial action. The FW Act does not expressly prohibit ‘unprotected industrial action’ except, for example, during the currency of an existing enterprise agreement (s 417). ⁵ But there are three main consequences of getting protected industrial action wrong.

First, the immunities cease to apply. This means that employers can seek injunctive relief in the courts, pursue unions for the losses attendant on industrial action and take disciplinary action against employees.

Second, the FWC is required to make an order that unprotected industrial action stop, not occur, or not be organised (s 418). Breach of an order under s 418 (technical, inadvertent or otherwise) can lead to the imposition of penalties in the Federal Court. Since the High Court decision in Esso, it has also meant that the employees cannot take further protected industrial action to support their claims in bargaining for that agreement.

Third, it was held in Esso that where a bargaining representative intends to take protected industrial action but fails to comply with the prerequisites, the unprotected industrial action will breach the prohibition in FW Act s 343 on action taken with intent to coerce an employer in respect of the making of an enterprise agreement (Esso at 425). In other words, if parties try to take lawful industrial action and get it wrong, the otherwise entirely lawful intention of the bargaining representative to coerce the employer to agree to their claims in bargaining translates into an unlawful intention to coerce the employer in breach of the FW Act. Inadvertent and technical failures to comply with the industrial action provisions are now effectively unlawful.

Coupled with these potential consequences is the significantly enhanced role of enforcement agencies in policing compliance with the FW Act. While employers can pursue remedies at common law and under the FW Act, enforcement proceedings may also be initiated by the Fair Work Ombudsman, which has authority under the FW Act to investigate and bring proceedings in respect of contraventions of the Act including for unlawful industrial action or for non-compliance with FWC orders in respect of industrial action (s 682). ⁶ Hardy (2018) observes that in the past 25 years our industrial relations system has changed from one where enforcement proceedings were relatively uncommon to one where highly active regulators operate alongside employers who are more willing than ever to bring actions both under the common law and in respect of breaches of the statutory regime (p. 2018: 204).

Why is it so hard for workers in Australia to take industrial action when they need to?

In the lead up to the 2007 federal election, the Australian Labor Party (ALP) faced a campaign from the Coalition seeking to paint them as soft on strikes, and their election as likely to lead to an avalanche of industrial action (Muir, 2008: 160–161). The ALP’s response, promising ‘clear tough rules’ on industrial action (Rudd and Gillard, 2007) meant that when they came to developing the legislative regime that would become the FW Act, too many of the preexisting provisions regulating industrial action were retained (McCrystal, 2009). The regime was not sufficiently overhauled to resolve many of the problems identified in this discussion which date back to before it was enacted.

A combination of restrictive judicial interpretation and statutory amendments since that time has further narrowed the scope of lawful industrial action under the FW Act.

The use of an immunity protected industrial action has left the legislative provisions vulnerable to formalistic interpretations by a judiciary ultimately seeking to protect the common law. Furthermore, the scope of protections for workers in the exercise of industrial activities (including protected industrial action) is cause for great concern, with the decision in BHP Coal leaving them substantially devoid of content.

Initiation of bargaining under the FW Act is controlled by employers and the FWC, with substantive agency in this respect having been removed from workers in 2015 when their capacity to use industrial action to force employers to the bargaining table was removed. Once the window for industrial action opens, the reasons for which protected industrial action can be taken are few, and the claims that can be supported are not always entirely clear. Break an order of the FWC or a court for any reason and protected industrial action ceases to be available.

The prerequisites to industrial action are complex and administratively demanding. Superficially reasonable requirements for ballots to approve proposed action have been implemented in a way that undermines the ability of those members to approve the action. Unreasonable quorum requirements and lack of accessibility to attendance and electronic balloting make the process more difficult. At almost every stage along the way there are opportunities for employers to intervene, imposing additional delay and complexity.

If the process is successfully navigated, effective industrial action that has an economic impact or threatens the welfare of third parties can be stopped by the FWC with relative ease.

Finally, get it wrong and irrespective of one’s intention to comply with the laws, threatened or actual unprotected industrial action will breach the prohibition against coercion and leave those involved potentially subject to civil penalties in the Federal Court.

This overview of the legal framework for lawful industrial action under the FW Act paints a grim picture in which the total impact on the capacity of workers to take lawful industrial action is much greater than the sum of each part. The system is rife with complexities and some perplexing rules, which make compliance difficult even for the most seasoned industrial campaigners (e.g. the AWU in the Esso case).

The real question appears not to be ‘Why is it so hard for workers to take industrial action when they need to?’, but Why do they continue to try to act lawfully at all?’. The introduction of the protected industrial action regime in 1993 created a ‘zone’ of lawful industrial action, which had the converse effect of more clearly delineating when industrial action was unlawful and unacceptable. In the subsequent 25 years, most unions in Australia have tried to play by the book, seeking to operate within the sphere of legality. However, the rule book keeps changing, the scope of action narrowing, the resource and administrative burdens increasing, and the stakes, in terms of getting it wrong, rising.

Since the decision of the High Court in Esso, it has been reasonably clear that the burden of trying to meet the protected industrial action provisions may now be so high that the price of getting it wrong – of simply abandoning the attempt to act legally – may be worth paying. Sally McManus’s comments about unjust laws demonstrate that given the degree to which the legal rules are stacked against workers, rejecting the legal course of action in favour of unlawful strikes is becoming more appealing.

However, the potential impact of such a course of action is unclear. The last time that all industrial action in Australia was effectively unlawful was before the introduction of protected industrial action in 1993. While the legislation at the time ‘bristled’ with enforcement tools, in practice they were largely unused (Creighton, 1991). The enforcement context for unlawful industrial action has now changed significantly, with both employers and well-resourced enforcement agencies more prepared to use the law to control industrial conflict. Even if the costs associated with compliance are very high, the potential economic consequences of a sustained campaign of unlawful industrial action could ultimately cripple a union.

If workers are to be able to take industrial action when they need to, the system must be reworked to ensure that this is possible. The impact of the laws needs to be considered in their totality to ensure that the sum of the regulatory system does not add up to more than its parts, undermining the ability to take protected industrial action rather than facilitating it. The most pressing changes in this respect are to ensure that unintentional or minor deficiencies in the processes accompanying proposed industrial action do not result in a loss of protection, and that any loss of protection does not invariably mean commission of the offence of coercion. In the longer term, a change of mindset on the part of regulators is required to embrace industrial action as a necessary component of our industrial system. If we are to maintain a system of voluntary collective bargaining in which the interests of workers are respected and protected, the legislation must permit robust bargaining supported by timely and straightforward access to industrial action for employees. Otherwise we are likely to see continued wage stagnation, accompanied by increased unlawful industrial action as unions and their members struggle with our unreasonable laws.