Proactive Antitrust: Expanding the Australian Regulatory Toolbox to Manage Algorithmic Pricing Risks

Alice Sharma, Freya Phillips Scholar, 2024 World Bank & IMF Fellow

Executive Summary

This paper tackles the risk of algorithmic collusion, in which pricing algorithms learn to coordinate prices without explicit agreement, thereby weakening competition and evading the reach of current competition law. The recommendation is to adapt a proactive antitrust strategy by introducing a new ACCC power to conduct market investigations before harm occurs. If structural features of a market are found to support tacit collusion, the ACCC could apply targeted remedies to restore competitive pressure. Thus, the focus would shift from attributing liability to market design. While there are costs to regulators and firms, similar regimes have been implemented in foreign jurisdictions and yielded long-term benefits to consumers. This policy would bring Australia’s competition regime in line with the realities of algorithm-driven markets.

Problem Identification

The exponential advances in algorithmic technology have facilitated opportunities for market efficiency, particularly in the field of product pricing. Firms are now able to dynamically price their products using software that aggregates and analyses sales data to arrive at a profit-maximising price. However, there is an equal opportunity for exploitation. 

A study spanning 15 years of unleaded petrol prices in Perth identified patterns of unspoken collusion among fuel companies to coordinate prices and soften competition (Byrne & de Roos, 2015). The use of pricing algorithms, which may have learnt to implement collusive strategies with a finite phase of punishment followed by a return to cooperation, allowed firms to maintain supra-competitive price levels to the detriment of consumers (Calvano et al., 2018). More recently, a 2024 study found that a Large Language Model program directed to ‘seek long run profits’, a legitimate business objective, had consistently achieved that said objective by engaging in collusive conduct (Fish et al., 2024).

Unlike explicit collusion which is clearly prohibited by Section 45(1)(c) of the Competition and Consumer Act 2010 (Cth), tacit collusion achieved through algorithmic interactions may bypass traditional legal scrutiny given the absence of any explicit ‘agreement’. Thus, Australians face a dual threat. First, that in undermining fair competition, algorithmic collusion results in higher prices. Second, that the enforcement of existing competition laws, which were largely designed to capture human collusion, becomes largely toothless (Ezrachi & Stucke, 2016).

Context

In competitive markets, each economic operator must independently determine the price it intends to adopt. Collusion occurs when firms mutually coordinate their actions to manipulate markets, allowing them to maintain supracompetitive prices and limit competition (Posner & Easterbrook, 1981). This generally involves a communicated agreement between competitors, which is prohibited under Part IV of the Competition and Consumer Act 2010 (Cth).

Pricing algorithms may facilitate collusion but circumvent a communicated agreement in two ways. 

1. Hub and Spoke

   
   

   Firms often outsource their pricing to an intermediary algorithm provider to optimise pricing strategies. This delegation creates conditions for a hub-and-spoke-like framework to emerge.  In this scenario, the algorithmic service provider (the ‘hub’) centralizes pricing decisions for multiple competing firms (the ‘spokes’) (Harrington Jr, 2025). As illustrated in Figure 1, each spoke provides the hub with data input (including cost, product and inventory data), which is then analysed by the algorithm to determine market conditions. In doing so, it is able to effectively and efficiently determine the optimal price for each spoke’s product. Where this turns harmful is when the spokes’ reliance on the algorithm results in coordinated and thereby supra-competitive pricing. 

   
   

   

   Figure 1: Hub and Spoke Model

   The critical element that protects these arrangements from the grasp of antitrust law is that the spokes do not interact with each other, but it is their shared reliance on the hub that facilitates collusive outcomes . A troubling case occurred in the Rotterdam petrol market, where the hub of “a2i systems” was used to determine prices. When tested against a control group, it was found that the spokes averaged 5% higher profits, demonstrating the unintentional collusion by the petrol stations. The hub and spoke relationship here fell outside the scope of antitrust law since there was no agreed aim to coordinate prices to facilitate collusion (Schechner, 2017).

   
   

   Extending the analogy further, in circumstances where the horizontal competitors can be said to engage in the questionable conduct with the requisite degree of awareness and commitment, the ‘rim’ of the wheel will be complete and would fall foul of existing competition law. This was the case in the 2016 Eturas case adjudicated by the Court of Justice of the European Union (CJEU), where several travel agencies were the spokes connected to the hub of Eturas, an online travel booking platform. The CJEU held that the spokes’ knowledge and acceptance of a shared pricing structure facilitating collusion was indicative of their intention to collude, and could thereby attract liability (Court of Justice, European Union). 
   

2. Autonomous Tacit Collusion

   Here, despite competitors using independent pricing algorithms, the ability of the algorithms to react to competitors’ prices in a transparent market induces parallel prices movements (Agarwal, 2021). As seen in Figure 2, FIRM 1 employs ALGORITHM A to monitor the pricing of FIRM 2, reacting to price changes with a strategy to maximise profits. Through signalling, algorithms become aware of the interdependence of the two firms. As such, they recognise that they can profit by not only aligning price, but maximise profit by acceding to a mutual, higher price (Ezrachi & Stucke, 2016).

   
   

   

   Figure 2: Autonomous Tacit Collusion Model

   This behaviour is enhanced by the speed of pricing algorithms. For a hypothetical, imagine FIRM 2’s pricing ALGORITHM B is coded to simply match the price of ALGORITHM A. Any reductions in ALGORITHM A’s prices will therefore yield no benefits, since FIRM 2 will retaliate by immediately also reducing its price. Therefore, FIRM 1’s reduction will capture no new customers. Thus, ALGORITHM A is incentivised to set a cooperative equilibrium at a higher price, knowing it will be matched, and thereby enabling both FIRM 1 & 2 to increase profit (Capbianco et al., 2017). This is arguably more dangerous than traditional cartel conduct, because prices set will be more durable (Mehra, 2016). 

   
   

   A similar case occurred in the German fuel market. The Bundeskartellamt (Germany’s federal competition bureau), had begun an inquiry under § 32 e of the German Act against Restraints of Competition (ARC) into the retail fuel market, which it suspected was engaging in collusive behaviour. In a well-intentioned but ill-fated attempt to promote competition, the Bundeskartellamt required fuel companies to report any price changes in real time (Bundeskartellamt, 2011). Hypothetically, this would reduce consumers' search costs for finding cheaper products and force each firm to compete on price. Unfortunately, a study found that this transparency increased petrol prices by 1.2-2.3 euro cents and diesel prices by 2 euro cents (Ezrachi & Stucke, 2020). This has further been supported by simulation studies, which find that algorithms learn to initiate supra-competitive prices when instructed only to maximise profits (Calvano et al., 2018). Again, this fell outside the scope of antitrust law because there was no communication between firms. Instead, collusion was tacit, having been enhanced by a transparent market and speedy pricing algorithms that ensured discounting would bring no profits. As such, prices rose (Ezrachi & Stucke, 2020). 

   
   

Applied to Australian Law

Section 45(1)(c) of the Competition and Consumer Act 2010 (Cth) is the current source of liability. It provides that a corporation must not  “engage with one or more persons in a concerted practice that has the purpose, or has or is likely to have the effect, of substantially lessening competition (‘SLC’)”. There are two issues relevant to algorithmic pricing.

First, the ambiguity of what demarcates a ‘concerted practice’. There is no definition in the Act. The ACCC guidelines for the Act point to the judicial interpretations of foreign jurisdictions’ including the European Union, which defines it as “any form of coordination between undertakings” that doesn’t necessarily reach the threshold of an agreement but nevertheless “knowingly substitutes for the risks of competitive practical cooperation between them” (Australian Competition and Consumer Commission, 2016). European case law thus suggests that reciprocal communication between the coordinating firms is a relevant indicia for ‘concerted practice’. 

Regarding the reciprocal element, The Explanatory Memorandum to s 45 affirms that unilateral conduct falls outside the scope of the provision (Explanatory Memorandum, Competition and Consumer Amendment (Competition Policy Review) Bill 2017 (Cth), p. 29) . A case of autonomous tacit collusion as discussed prior, where only FIRM 1 makes the decision to increase its price knowing its competitor’s behaviour (recall that ALGORITHM 2’s strategy is to always match), would thus not satisfy the ‘reciprocal’ requirement. The communication element also raises difficulties. Traditional antitrust law has evolved around human actors, where proving collusion relies on evidence of a 'meeting of minds' or mutual cooperation (Agarwal, 2021). Algorithms, operating without conscious decision-making, make it difficult to prove the existence of communication in this sense. 

The second issue that arises is the SLC test. Even if collusive behaviour is found, there is no liability unless it can be shown to substantially lessen competition. This sets a high threshold. It is also an unclear threshold since “no one knows with any clarity what ‘substantial’ means in the SLC test” (Nicholls & Fisse, 2018).  In hub-and-spoke scenarios, where a platform operator or central firm communicates with multiple users, it may still be possible to demonstrate an SLC if the spokes consciously align their behaviour. However, in the case of autonomous tacit collusion where algorithms independently pursue collusive outcomes, establishing a causal link becomes much harder. In practice, where algorithmic pricing is widespread and constantly evolving, constructing a credible counterfactual is likely to be highly challenging,

Despite the broader drafting of section 45, these evidentiary hurdles mean successfully establishing liability for algorithm-driven collusion remains difficult.

Examining the Regulatory Toolbox

A cursory look over Australia's regulatory ‘toolbox’ (referring to the mechanisms which regulators have to intervene in the competitive landscape of the market) reveals a significant regulatory gap. Section 155(1) of the CCA outlines the investigatory powers of the ACCC, which may be exercised if the commission requires evidence relating to a “matter” that “constitutes or may constitute a contravention” of the law. As such, an investigation may only begin once an infringement has materialised. This limits the ability of the regulator to proactively gather information into the conditions of an industry, and largely places regulators on the back foot (Feiglin, 2020). 

The UK’s toolbox contains an arguably much more powerful tool, one which the EU implemented a watered down iteration of in 2022 (Kuipers and van Roosmalen, 2024). In broad terms, this tool allows the regulator to circumvent the otherwise tedious process of determining an infringement of law and instead enables them to impose regulatory remedies ex ante (before the fact). Under Chapter 4 of the Enterprise Act 2002, the Competition and Markets Authority (CMA) is able to conduct independent investigations to determine features that distort competition within an industry, and then implement structural remedies designed to restore competition (Ralston-Smith & Johnson, 2020). By way of example, the 2009 investigation into the UK airports revealed a severe lack of competition materialising in poor passenger experience (Clarke et al., 2009). After prescribing the industry various divestiture measures and price controls, airports experienced an increase in traffic, improved connectivity and choice, and downward pressure on fares. Overall, the benefits were estimated to be £870m from 2016-2020 (Competition and Markets Authority, 2016). It is important to note that this tool focuses on industry-wide practice and structures rather than individual liabilities.

Policy Options

Option 1: Amendments to Existing Competition Law If the policy goal is legal — that is, to ensure that improper use of pricing algorithms attracts liability — then legislative reform is required. 

Two key amendments to Section 45(1)(c) of the Competition and Consumer Act 2010 (Cth)  are proposed: 

1. Define ‘concerted practice’ and clarify that signalling or algorithmic communication can constitute coordination 

   The EU’s definition of concerted practice, which is already referenced in the guidelines for the Act, should be codified into the CCA. A legislative amendment could also clarify that coordination falling short of an agreement, including: algorithmic signalling, mutual responsiveness, or sharing of pricing data via a common hub, may constitute a concerted practice. This would mean that in the case of autonomous tacit collusion, where ALGORITHM A responds to ALGORITHM B’s unilateral price hikes by fixing a supracompetitive price, the court could infer coordination from structured market responses. Similarly, in a hub-and-spoke case, if the hub disseminates non-public pricing information and the spokes adopt parallel pricing strategies, then awareness or reasonable foreseeability of that information transfer could establish coordination. 
   

2. Lower the threshold from a ‘substantial’ to ‘any demonstrable lessening of competition’  Given that courts have shown a reluctance to find liability where the effect on competition is marginal, the SLC threshold should be lowered to ‘any demonstrable lessening of competition’. This would better reflect the dynamic harms caused by algorithmic collusion. Taken together, these reforms would shift the legal presumption: from requiring the ACCC to prove a human-led, intentional agreement with clear causal effects, to allowing courts to infer coordination from structured market behaviour and algorithmic responsiveness and thus closing the enforcement gap left open by current doctrine.

   
   

   However, the extent to which increased legislative drafting will help capturing algorithmic collusion is unclear. It may be the case that explicating the bounds of the law too early in the development of artificial intelligence  technologies will create a rigidity in the law that will render it unresponsive to technological changes. Additionally, too broad of a definition may blur the line between collusion and parallel conduct such that it may capture instances of legitimate conduct of firms. This overregulation may potentially hinder innovation, as firms may avoid deploying competitive algorithms for fear of repercussions. It may also make it undesirable for global firms to operate in Australian markets

   
   

   Further, this proposal would place the burden on the judicial branch and regulatory agencies to prosecute collusive outcomes, rather than addressing the structural incentives that give rise to these economic behaviours in the first place. In this way, the focus of policy liability attribution instead of promoting a competitive market.

Option 2: Auditing Pricing Algorithms Prior to Market Use

As such, Option 2 focuses on how to minimise collusive outcomes by implementing a robust regulatory framework that requires algorithms to pass an audit before they are deployed into markets. The focus thus shifts to economic outcomes.  

There are 3 main parts:

1. Programming Guidelines The first step would involve the ACCC issuing programming guidelines to various sectors targeting specific algorithmic strategies known to facilitate coordination. In this way, a compliance by design strategy is employed that places the initial burden on firms (Harrington Jr, 2017).
   

2. Increase Data Discovery Powers of the ACCC A ‘data request power’ would then allow the ACCC to obtain data from firms in real time. This would require a legislative expansion of the commission’s investigatory powers under section 155 of the CCA. The aim of collecting such data is to enable the regulatory body to develop and continually refine its auditing mechanism (Feiglin, 2020). Although the exact details are technically complex and outside of the scope of this paper, it should be noted that data access at this scale raises significant confidentiality risks. It must be legislated in a way that improves information access  without very real risk of overreach. 
   

3. Audit The final step involves requiring firms to submit pricing algorithms for audit prior to deployment. In theory, this ex-ante review would act as a safeguard, ensuring algorithms are not designed in ways that facilitate collusion. However, pre-deployment auditing faces serious limitations. Many modern algorithms, particularly those driven by machine learning, operate opaquely and adapt based on real-time market inputs, behaviour that is often unpredictable in test environments (Ezrachi & Stucke, 2017).

   
   

   Nonetheless, emerging research in software engineering and market analysis suggests some promise (Hartline et al., 2024). With tools designed to flag anomalous price patterns or departures from rational profit-maximising behaviour, regulators could identify moments that suggest tacit signalling or coordination. In such cases, the burden may shift to firms to demonstrate their algorithm’s behaviour was competitively neutral. However, even this ‘burden-shifting’ model depends on advances in detection technology that remain underdeveloped, and face significant challenges in scalability, interpretability, and robustness. Pre-approval processes also risk entrenching high compliance costs and raising barriers to entry, which could undermine competition in the very markets they aim to protect.

   
   

   Auditing remains the weakest link in the regulatory toolkit  - not due to lack of ambition, but because the technological and evidentiary ground is not yet solid enough to support large scale implementation.

   
   

Option 3: A New Australian Competition Tool

Alternatively, policy could empower regulators to intervene in markets that appear vulnerable to algorithmic collusion before that coordination materialises. Here the focus lies in not attributing liability to any one firm, but improving economic outcomes for markets. This proposal calls for a section to be introduced into the CCA enabling powers to, following a market investigation, intervene in markets by imposing remedies that are focussed on correcting structural barriers to competition. As such, the regulatory toolbox is expanded to favour proactive antitrust. Policy design would closely resemble that of the UK’s new competition tool as discussed earlier. 

It is important to note the risk that such wide market powers may give rise to an unacceptable level of interventionism, which would have the adverse effect of lessening competition in markets. There are arguments from industry that such interventions would be disproportionate to harm, especially considering that this tool is available in circumstances without an infringement of the law. Such a power then must be designed with strong procedural checks and balances that prevent overreach, including potentially a clause that requires the regulatory body to ensure that there is no equally effective remedy available that is less intrusive. 

Policy Recommendation

To effectively address algorithmic collusion, a successful policy must: 

1. Prevent anti-competitive conduct arising from pricing algorithms.

2. Enhance the ability of regulators to detect and address algorithmic collusion early.

3. Avoid stifling innovation or deterring global firms from operating in Australian markets.

4. Ensure proportionality by targeting problematic actors and behaviours without imposing unnecessary burdens on legitimate business practices.

Option 3 focuses on economic outcomes coupled with its ability to ensure timely and effective intervention fulfils this criteria. It is the recommended policy. 

Ex Ante Market Investigations The first limb is an investigatory power. The ACCC would be able to launch a formal market investigation where there is a reasonable suspicion that features of a market are producing or facilitating tacit collusion, even in the absence of a proven legal infringement. This may be an observed pattern: prices consistently higher than competitive benchmarks, algorithmic behaviour that converges without explicit agreement, or repeated parallel conduct in markets characterised by high transparency.

It is important to legislate a time limit for such an investigation, not only for the purposes of ensuring timely intervention, but also to provide firms with certainty. Similar to the UK model, the initial assessment should be limited to 12 months, and a further 6 months should be devoted to designing remedies. There should also be an appeals option as a check to ACCC overreach.

Figure 3: Proposed Timeline

This power is ex ante and diagnostic in nature. It reflects a shift in regulatory philosophy from investigating after  competition has failed, to proactively identifying markets at risk of failing. The point is to recognise that in algorithmically driven, oligopolistic markets, collusion can emerge endogenously from the structure itself. These are not always cases of bad actors. They are cases of bad dynamics and our current legal framework offers no way to address them until after harm has crystallised. A market investigation tool would close that gap.

Market Based Remedies

The second element of the proposal is a remedial power. If, following an investigation, the ACCC concludes that certain market features are conducive to sustained collusion or dampened competition, it could propose proportionate interventions designed to disrupt those dynamics and restore competitive pressure. Crucially, these remedies would be structural, not punitive. Firms would not be sanctioned for their conduct; rather, the market itself would be recalibrated to reduce the likelihood of collusive outcomes (Ralston-Smith and Johnson, 2020).. Such remedies could include limiting public price transparency (to reduce algorithmic price-matching), requiring staggered pricing updates (to prevent simultaneous adjustments), mandating algorithmic explainability and documentation, or even restricting the use of certain pricing algorithms in sensitive markets. 

The remedies would be targeted and subject to oversight. Their aim would not be to prevent technological innovation, but to ensure that innovations do not undermine the foundations of a competitive market. Limits in the legislation would take into account clear economies of scale that justify market intervention. It is important that we emphasise a co-regulatory approach, including mechanisms for industry wide consultation that ensure workable and proportional solutions.

Costs

Taxpayers (who fund the ACCC) and firms (who are the subject of ACCC intervention) are likely to incur a direct cost. However, the long term monetary and structural benefits of improving competition in Australian markets are significant enough to warrant such spending. In the UK market for instance, estimations from the CMA found that from 2016-2019, the total direct consumer benefits from its interventions through the market study/investigation regime were £2.5 billion GBP, equivalent to $5.2 billion AUD (Competition and Markets Authority, 2019). 

Risks

While this proposal aims to strengthen competition and protect consumers from algorithmic coordination, it introduces real risks that must be acknowledged. Expanding regulatory powers in the absence of legal infringement may be seen as overreach, potentially attracting political resistance or undermining business confidence. Australia's competition regime has traditionally been reactive; a shift to proactive, ex ante intervention marks a significant institutional change. There is also a risk that smaller firms, with limited resources, may be disproportionately burdened by compliance requirements such as mandated algorithmic explainability or pricing update restrictions. Poorly calibrated interventions could entrench incumbents rather than improve competition. Similarly, there is a danger of false positives: not all price convergence is collusion. Overzealous regulation could suppress efficient market behaviour and deter innovation. Global tech firms may view the uncertainty of being investigated without clear wrongdoing as a deterrent to investing or operating in Australia, particularly absent international alignment. 

However, these risks can be mitigated. Built-in safeguards such as strict time limits, appeal mechanisms, industry consultation, and a requirement to consider less intrusive options may help to ensure that interventions remain proportionate and economically justified. While no policy is risk-free, the dangers of inaction in rising prices, ineffective enforcement, and systemic algorithmic collusion pose a far greater long-term threat to competitive markets.

References

Agarwal, A. (2021). Price setting algorithms and collusion: An Australian perspective. Competition and Consumer Law Journal, 28, 316. LexisNexis. https://plus.lexis.com

Australian Competition & Consumer Commission. (2016, September). Framework for concerted practices guidelines. https://consultation.accc.gov.au/legal-economic/draft-framework-for-concerted-practices-guidelines/

Bundeskartellamt. (2011). Fuel Sector Inquiry — Final Report. https://www.bundeskartellamt.de/SharedDocs/Publikation/EN/SectorInquiries/Fuel%20Sector%20Inquiry%20-%20Final%20Report.pdf

Byrne, D. P., & de Roos, N. (2015). Learning to coordinate: A study in retail gasoline. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2570637

Calvano, E., Calzolari, G., Denicolo, V., & Pastorello, S. (2018). Artificial intelligence, algorithmic pricing and collusion. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3304991

Capobianco, A., Gonzaga, P., & Nyeső, A. (2017). Algorithms and collusion - Background note by the Secretariat. Organisation for Economic Co-operation and Development. https://one.oecd.org/document/DAF/COMP(2017)4/en/pdf

Clarke, C., Carstensen, L., Collings, J., Haskel, J., Holroyd, R., & Moizer, P. (2009). BAA airports market investigation: A report on the supply of airport services by BAA in the UK. Report prepared for the Competition Commission.

Competition and Markets Authority. (2016). BAA airports: Evaluation of the Competition Commission’s 2009 market investigation remedies. Report prepared for the Competition and Markets Authority.

Competition and Markets Authority. (2019). CMA impact assessment 2018/19. https://assets.publishing.service.gov.uk/media/5edf6607d3bf7f12eb61a92d/CMA_Impact_Assessment_Report_2018_19_Final2.pdf

Court of Justice of the European Union. (2016). “Eturas” UAB and Others v Lietuvos Respublikos konkurencijos taryba (C-74/14). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A62014CJ0074

Ezrachi, A., & Stucke, M. E. (2016). Virtual competition: The promise and perils of the algorithm-driven economy. Harvard University Press.

Ezrachi, A., & Stucke, M. E. (2017). Two artificial neural networks meet in an online hub and change the future (of competition, market dynamics and society). SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2949434

Ezrachi, A., & Stucke, M. E. (2020). Sustainable and unchallenged algorithmic tacit collusion. Northwestern Journal of Technology and Intellectual Property, 17(2), 217.

Feiglin, N. (2020). Algorithmic collusion and scrutiny: Examining the role of the ACCC’s information gathering powers in the digital era. University of New South Wales Law Journal. https://doi.org/10.53637/ecwn8597

Fish, S., Gonczarowski, Y. A., & Shorrer, R. I. (2024). Algorithmic collusion by large language models. arXiv preprint. https://arxiv.org/abs/2404.00806

Harrington Jr, J. E. (2017). Developing competition law for collusion by autonomous price-setting agents. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3037818

Harrington Jr, J. E. (2025). Hub-and-spoke collusion with a third-party pricing algorithm. SSRN Electronic Journal. https://ssrn.com/abstract=5010894

Hartline, J. D., Long, S., & Zhang, C. (2024). Regulation of algorithmic collusion. In Symposium on Computer Science and Law (CSLAW ’24) (pp. 1–11). ACM. https://doi.org/10.1145/3614407.3643706

Kuipers, P., & van Roosmalen, J. (2024). The new competition tool: What is it and why do national regulators want it? Bird & Bird. https://competitionlawinsights.twobirds.com/post/102jj79/the-new-competition-tool-what-is-it-and-why-do-national-regulators-want-it

Mehra, S. K. (2016). Antitrust and the robo-seller: Competition in the time of algorithms. SSRN Electronic Journal. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2576341

Nicholls, R., & Fisse, B. (2018). Concerted practices and algorithmic coordination: Does the new Australian law compute? Competition and Consumer Law Journal, 26(1), 82–102. https://search.informit.org/doi/10.3316/agispt.20181016002845

Posner, R. A., & Easterbrook, F. H. (1981). Antitrust cases, economic notes, and other materials. West Academic Publishing.

Ralston-Smith, H., & Johnson, M. (2020). What problems can the European Commission’s New Competition Tool fix? Oxera. https://www.oxera.com/insights/agenda/articles/what-problems-can-the-european-commissions-new-competition-tool-fix/

Schechner, S. (2017, May 8). Why do gas station prices constantly change? Blame the algorithm. The Wall Street Journal. https://www.wsj.com/articles/why-do-gas-station-prices-constantly-change-blame-the-algorithm-1494262674

Legal Citation

Explanatory Memorandum, Competition and Consumer Amendment (Competition Policy Review) Bill 2017 (Cth). https://parlinfo.aph.gov.au/parlInfo/download/legislation/ems/r5851_ems_0b6ffc49-7398-409a-8e46-4873853a475f/upload_pdf/625422.pdf

Statute

Competition and Consumer Act 2010 (Cth) ss 45(1)(c), 155(1)

The views and opinions expressed by Global Voices Fellows do not necessarily reflect those of the organisation or its staff.