News feeds are no longer free: policy implications from Australia

ABSTRACT In 2021, the Australian Government introduced its News Media and Digital Platforms Mandatory Bargaining Code (‘the Code’) to support the sustainability of the country’s journalism. Using Event Study Methodology, we find a collectively positive stock market reaction in the news media industry to the introduction of the Code, although different patterns amongst individual firms and groups by size to the different stages of the Code’s development. The large-firm group records larger gain in terms of equity value, but a lower percentage gain than small and medium enterprises. The economic importance of our findings reinforces the Australian Government’s regulatory approach as a means to sustain the legitimate interest of businesses and consumers in the news media sector.


I. Introduction
Digital platforms are now central to the development of the global economy and society.Their ubiquity in the social dynamics transforms the way people access information and connect with others.Digital platforms in business enhance competitiveness and productivity.However, the novelty in the business model of online platforms and their dominant position in certain economic sectors may disrupt traditional business models and raise concerns over competition risks (OECD 2021a).
To support the adaptation to digitalization of businesses, Australia, along with many other jurisdictions around the world, evaluated digital disruption to the media sector and its effects on fair competition.Under the direction of the Australian Government, the Australian Competition and Consumer Commission (ACCC) published its Digital Platforms Inquiry report (DPI report) in 2019, which comprehensively assessed the impact of digital platforms (Google and Meta 2 ) on competition in the media and advertising services markets.The DPI report (ACCC 2019) finds that digital platforms have market power over their news media counterparts, based upon the value offered to users and the lack of close alternatives in the digital markets.Whilst Google and Meta are unavoidable partners for many news media businesses as they offer critical sources of referrals and avenues to reach potential customers, no single media business is necessarily the only choice of partner for these digital platforms.The report also reveals that Google and Meta alone captured approximately two-thirds of the Australian online advertising market in 2018, accompanied by a number of media business closures and journalism job losses between 2008 and 2018. 3As a result, the report also shows concerns over the underprovision of public interest journalism. 4On this basis, the Australian Government introduced the News Media and Digital Platforms Mandatory Bargaining Code ('the Code') on 25 February 2021 to address the bargaining power imbalances between digital platforms and the news media businesses.
Efforts to motivate digital platforms to compensate news publishers are not new.These digital giants are under more scrutiny worldwide (Brevini and Swiatek 2020).In 2014, Spain passed an Intellectual Property law which required online content aggregators such as Google News to pay a mandatory monthly fee to the organization representing Spanish newspapers, the Association of Editors of Spanish Dailies.Google responded by shutting down its news site.A 2019 Copyright law in the European Union also shifts more power to news publishers by setting copyright as a legal basis for compensation negotiations between presses and online platforms (EU 2019).While these countries seek to address similar issues in regulating the influence of Big Tech companies, they do so through the lens of intellectual property policy.With the introduction of the Code, the Australian government is the first country in the world to govern the relationship between digital platforms and news media businesses through an antitrust approach.Australian Treasurer Frydenberg (2020) describes the Code as the world's historical attempt to 'create a level playing field where market power is not misused and there is appropriate compensation for the production of original news content', and also 'to ensure a sustainable and viable Australian media landscape'.
The applicability of the Code motivates us to examine if it sends a positive signal to the news media industry stock market through improved investor confidence.The economic importance of this matter is twofold.On the one hand, Australia is not alone in their battle to regulate Big Tech companies, given the latter's global scale of influence.A number of countries around the world are considering to follow suit, including the US, the UK, India, New Zealand, to name just a few.Particularly, Canada recently published its Online News Act in April 2022 following key provisions of the Australian Code.On the other hand, Google and Meta are also not alone in the race of novel technology.A number of emerging digital platforms like TikTok shall take over Google and Meta in no time in offering diverse options to access news for consumers. 5Our findings shall, therefore, address the urgent quest by policymakers around the world who are considering the Australian approach as a potential policy choice.
Whilst the Australian Code has drawn increasing scrutiny from regulators, it has also triggered interest amongst scholars.A number of papers looked into issues around the Australian media landscape in general and particularly the Code during 2020 and 2022.This literature advances theoretical propositions and policy recommendations (Hess 2020;Hess and Waller 2021;Brevini 2021;Bruns and Angus 2021;Leaver 2021) or constructs descriptive analyses of interviews around the Code with relevant stakeholders (Meese and Hurcombe 2021), or reviews the actual deals scored following the introduction of the Code (Bruns 2021); or induces a qualitative assessment of the Code through public responses (Bossio et al. 2022).To our knowledge, there is yet any paper providing empirical evidence on the impact of the Code.Thus, our paper complements existing literature by offering such evidence through examination of stock market performance of the news media businesses following the introduction of the Code.
In our research, we use Event Study Methodology to examine abnormal returns (AR) and cumulative abnormal returns (CAR) of a set of securities of Australian news media companies around different development stages of the Code.As AR and CAR, by definition, reflect the difference between the actual return of a security and the expected return, the empirical results reflect the changes in the financial performance of news media firms in the stock market throughout the development of the Code.We find a collectively positive stock market reaction of the news media industry in Australia to the introduction of the Code, albeit inconsistent patterns amongst individual firms and groups by size to the different stages of the Code's development.While all media companies, regardless of size, record sizable monetary gains following the introduction of the Code, we observe a huge gap in the equity gains between large firms and firms of small and medium size.A similar pattern of CAR results obtained from different estimation models: Market Model (MM), Market Adjusted Model (MAM), and Carhart Four-Factor Model (CFFM) as well as a null result from a placebo test, reaffirms the robustness of our empirical findings.
From an applied perspective, our research provides evidence-based rationales of potential regulatory outcomes for policymakers and regulators.Because governments around the world are watching the practicality of the Australia's regulatory approach in dealing with the increasing dominance of digital platforms, our findings reinforce the empirical basis for regulatory reform that sustains the legitimate interest of the businesses and consumers in the news media sector.Furthermore, when considering a similar approach to regulate Big Tech companies, governments should take into account the current state of the country's media landscape.An inclusivity plan should be in place to ensure not only large players, but also small-and medium-sized businesses are equally legitimate beneficiaries of government policies, in the pursuit of a sustainably healthy and diverse media ecosystem.
We proceed as follows.Section II reviews background of the event.Section III discusses methodology and data.Section IV presents empirical findings.Section V conducts robustness tests.Section VI concludes.

Context of the Code
The ACCC published the DPI report in 2019, which comprehensively assessed the impact of digital platforms (Google and Meta) on competition in the media and advertising services markets in Australia.The report finds that Google and Meta, by means of the value offered to users and the lack of close alternatives, have more substantial market power over their news media counterparts in the digital markets.By offering critical sources of referrals and avenues to reach audience, Google and Meta become unavoidable partners for many news media businesses, whereas no single media business is necessarily the only choice of partner for these digital platforms.As such, digital platforms are found to possess more substantial bargaining power in their commercial dealings with individual news media businesses, which results in media businesses accepting less favourable commercial terms than they otherwise might accept relating to the availability of their content on each of these platforms.Such imbalances in the bargaining power affect the news media businesses' ability to monetize their news content and, ultimately, to fund original public interest journalism.In addition, the ACCC also partly attributes the collapse of the news media sector in Australia to the non-regulated increase in the use of the digital platforms.For example, 106 local and regional newspaper titles closed between 2008 and 2018, leaving 21 of Australia's 537 local government areas without coverage from a local newspaper (in either print or online formats); 3,000-3,200 journalism jobs were lost between 2012 and late-2017.In 2018, Australian online advertising expenditure was AU$8.8 billion, accounting for approximately 50% of total Australian advertising expenditure, while Google and Meta alone captured approximately two-thirds of that market (ACCC 2019).On this basis, the Australian Government directed the ACCC to develop a voluntary Code due by November 2020.The Government added a caveat that if negotiations did not progress by May 2020, it would develop other options, including potentially a mandatory Code.
In April 2020, the ACCC advised in its progress report that it was unlikely that any voluntary agreement on content payment would be reached.At the same time, the Australian media sector was under increased financial pressure due to a sharp decline in advertising revenue driven by the COVID-19 pandemic.The Australian Government, on 20 April 2020, directed the ACCC to develop a mandatory Code setting out steps that both news media companies and digital platforms are required to negotiate payments for news contents featured across those online platforms.
The draft Code was published for public consultation between May and August 2020.On 9 December 2020, the draft Code was officially introduced to Australian Parliament.Although it reflected many changes based on comments during consultation process, the draft Code still received negative reactions from Google and Meta -the targeted digital platforms.On 22 January 2021, Google threatened to shut down its search engine in Australia, while on 17 February 2021, approximately one week before the official passage of the Code, Meta temporarily blocked all news posts in its local site.In response, while remaining firm with key technical elements of the Code, the Australian Government did make some concessions in their final amendments to the Code announced on 23 February 2021.
Under the original proposals, Google and Meta were required to individually negotiate content prices with publishers within 3 months, or be forced into an arbitration process where a government-appointed panel chose between the publisher and tech giants' proposals.Under the final version, the government agreed to amendments that gave the tech giants greater control over the negotiation process in that the Code might not be applied to Meta or Google if they already had deals to pay news publishers for their content in place.The Australian government also gave platforms a month's notice, with forced arbitration a measure of last resort.There must also be a two-month period of mediation before arbitration.The final Code was then passed by Australian Parliament on 25 February 2021, with immediate effect from 3 March 2021.
The tension between news media and digital platforms for years has motivated regulatory interventions by governments across jurisdictions.The European Union and several member countries address this tension by adopting an EU Copyright, which uses copyright as the legal basis to make digital platforms pay for news content.The Australian Government, on the other hand, embraces an antitrust approach, which stems from the competitive dynamics between digital platforms and news publishers for user attention and advertising revenue.The market power of a few ad-funded digital platforms limits the online monetization of news contents and thus decreases the sustainability of public interest journalism, threatens the viability of news businesses, and reduces the quality and coverage of news, as well as exacerbates social harm caused by misinformation (OECD 2021b).Although sharing a similar aim to ensure fair compensation for news content, the Australian Code is unique in its antitrust approach, which aims to address the bargaining power imbalances between digital platforms and Australian news businesses.

Scope and intent of the Code
The Code supports the sustainability of the Australian news media sector by addressing bargaining power imbalances between digital platforms and Australian news businesses.It sets out six main elements: bargaining, compulsory arbitration, general requirement, non-differentiation requirements, contracting out and standard offers; as well as clear criteria to define a news business and a digital platform corporation (See Online Appendix A and B for details).
The Code is intended to impact the performance of news media businesses in two ways.First, the Code incentivizes commercial deals between digital platforms and news media businesses.This profit for the news media businesses could not have been achieved without the Code.The former Chair of the ACCC, Sims (2022), estimates an amount of over AU$200 million per annum is delivered to media sector from those deals.Sims estimates that the compensation amounts to approximately 20% of Australian journalists' salaries and likely more than 20% of eligible publishers' combined EBITDA. 6While neither Google nor Meta was designated under the Code, the threat of designation with the ultimate recourse of arbitration led both companies to engage in negotiations with publishers.The cost for failure to comply with the Code is not small -maximum AU$10 million or 10% of the Australian turnover during the last 12 months.
In fact, according to Treasury (2022), there were 34 deals scored by March 2022 (Google privy to 20 and Meta to 14).Further, there were 146 news businesses covered by those deals -representing 60% of print and digital news outlets in Australia (PIJI 2022).In addition to the upfront monetary profit, the Code is also reported to incentivize news media industry to expand their businesses.Recall from section 2.1 concerning a number of news firm closures and job losses between 2008 and 2018.One year since the passage of the Code, there were 45 expansions of news production, 11 new newsrooms, 11 service increases and 5 new print editions (PIJI 2022).Comparatively, in the 12 months leading up to the Code, 87 news outlets were founded, 2 newsrooms opened, 5 publications increased their service and 9 began a new print edition (PIJI 2022).Although the Code does not mandate companies to use the money gained from such commercial deals to reinvest directly in journalism, several media companies publicly link their business expansion as the direct result of such deals.For example, the Guardian claimed that revenue from the Code helped its building Australian newsroom from 0 to more than 100 journalists in 2021 (Grueskin 2022) or the ABC attributed its expansion by adding 55 new jobs, 10 new and 9 expanded bureaux as a direct result of revenues arising from their deals with digital platforms (PIJI 2022).In short, it is the mandatory nature of the negotiation for the news content payment provided by the Code that facilitates fair negotiations between digital platforms and news media firms, and hence incentivizing news providers to produce quality news.

The Australian media landscape
In the last decade, Australia is among the top countries in the world with the most highly concentrated media sector.Noam (2016) and the International Media Concentration Collaboration studied media ownership and concentration of 30 countries across the globe, and found Australia's media sector concentration in the top three in 2011, only after China and Egypt where the majority of media sector was owned by the government.Papandrea and Tiffen (2016) attributed the market dominance to News Corp, which controlled 57% of the newspaper market by circulation.This finding was reinforced by IBISWorld (2016), which found the total revenue of the three companies News Corp, Nine Entertainment and Seven West Media accounted for over 75% of the newspaper publishing industry revenue from 2015 to 2016.The research also found a high level of concentration for television and radio industry, in which the largest players accounted for 70% and 65% of the industry revenue respectively during 2015-2016.Brevini and Ward (2021) used measurements of readership, media ownership and revenue to assess media diversity in Australia, and found that News Corp had the dominant position, owning 59% share by readership, followed by Nine (23%), and Seven West (15%).In terms of revenue, News Corp once again was the market leader whose revenue in the newspaper market and commercial television market accounted for 62% and 40% respectively in 2020.News Corp, Nine and Seven West collected almost 80% of Australian free-to-air and subscription television revenues.In the radio market, News Corp, Southern Cross, Nine Radio and HT&E controlled almost 90% of the radio licences, and earned over 79% of total radio market revenues in 2020.
Thus, larger companies, which potentially possess higher corporate bargaining power (e.g. in terms of revenues, profits, resources for content creation, market campaign, policy advocacy, etc.) in comparison with smaller companies, are in a more advantageous position to strike more sizable deals with digital platforms (PIJI 2022).In fact, 248 news businesses, accounting for 296 news outlets, of which 96% are at local scale, are yet to have secured a deal to date (PIJI 2022).While the Code is intended to address the competition imbalance between digital platforms and news media industry, there is an uneven playing field between those news businesses with increased financial capacity from deals, and those without, leaving the latter at a distinct, competitive disadvantage, mostly amongst the smaller news producers.As part of this study, we empirically disentangle the impact of the Code on the financial performance of two groups of news media companies in Australia clustered by firm size: large-firm group and small-and medium-sized group of firms.If and to what extent such differences exist would imply the unintended effect of the Code in worsening the polarization of the media sector in Australia to the concentrated ownership of a few market dominants.

Methodology
We use Event Study Methodology (ESM) to examine the security price behaviour of selected news media businesses in Australia following the introduction of the Code.In practice, besides its major applicability in finance research, ESM has also been widely used in previous literature of different subsectors of economics including regulatory economics.For example, Schipper and Thompson (1983) investigate the impact of the merger-related regulations on the shareholders of acquiring firms; Binder (1985) examines 20 major regulatory changes which took place from 1887 to 1978 on stock returns; Prager (1992) studies the impact of the cable television deregulation during the time frame January 1981 and August 1985; Dnes et al. (1998) examine the stock market reaction to the price-capping measures of the UK electricity industry; Lamdin (1999) investigates the impact of the US ban on cigarette advertising on firm performance; Beber and Pagano (2013) examine the effects of the short-selling bans posed by regulators on stock prices between January 2008 and June 2009 in 30 countries; Lin et al. (2016) examine the impact the Chinese Politburo's announcement of an Eight-point Regulation initiating an extensive anti-corruption campaign on stock returns of all firms publicly listed on China Stock Exchange; Xiao and Gao (2017) examine the impact of China's Food Safety Law on Chinese listed firms in the food industry.

Event window
For regulatory events, the date that the new information actually reaches the securities market is not a concise and explicit one (Binder 1998).There is usually a long time-span encompassing an event window between the first and the last announcement of the regulation due to the nature of a policy making process, within which there will be multiple 'event periods' affecting the changes of the regulations, and constituting news to investors (Lamdin 2001).For example, the average time for approval of a regulatory change in Binder's study in 1985 is 18.5 months.In our case, the drafting of the Code itself took around 10 months to be completed (April 2020 to February 2021), including public consultation during May -August 2020, prior to the Parliament reading during December 2020 and February 2021.Table 1 shows key dates relating to the development of the Code.
For the purpose of conducting ESM, we identify three important dates that possibly send signals to news media investors in the stock market.The first date is Monday, 20 April 2020, when the Australian government announced that it had directed the ACCC to develop a mandatory Code to address bargaining power imbalances between Australian news media businesses and digital platforms.The intent of the Australian Government before this date had always been a voluntary mechanism whereby digital platforms and news media businesses could work out together a commercial deal for news content payment.In April 2020, the ACCC expressed their pessimism over the likelihood of any voluntary agreement could be met by businesses, while the pressure faced by the Australian media sector was exacerbated against The second date is Wednesday, 9 December 2020, when the draft Code was introduced to the Australian Parliament.This is a significant milestone in the law-making process in Australia, after an extensive period of pre-Parliament consultation process, marking one step closer to the official passage of the Code.While remaining firm with original intents and provisions, the draft Code introduced to Parliament also reflected several changes in response to stakeholders' submissions during consultation process, such as the acknowledgement of the value of traffic referred by the digital players to news media website from sharing their content during arbitration or adjustments to notifications around advanced warning of algorithm changes and data.
The third date is Tuesday, 23 February 2021, when the Australian Government made a public announcement on the final amendments to the draft Code, after a long debate the week before between the Government and digital platforms.Under the original proposals, Meta and Google were required to individually negotiate content prices with publishers within 3 months, or be forced into an arbitration process where a government-appointed panel chose between the publisher and tech giants' proposals.With these final amendments, digital platforms may not be subject to mandatory steps set out by the Code if they are deemed by the Government to have struck sufficient deals with news media companies.As the Code is unclear on how many deals are deemed sufficient, the practical designation of timeframe and online platforms subject to the mandatory implementation of the Code is also unclear.This creates uncertainty about the implications and practical implementation of the Code amongst stakeholders, including investors in the stock market.
Therefore, the full regulatory event window for our study spans over approximately 10 months, from 20 April 2020 to 23 February 2021, and consists of three event periods: event period 1 around 20 April 2020, event period 2 around 9 December 2020, and event period 3 around 23 February 2021.At each event period, we calculate AR and CAR of individual news media firms on the event date (T0), three-day sub-eventwindow T[−1,+1] including one trading day before and one trading day after the event date, and fiveday sub-event-window T[2-,+2] including two trading days before and two trading days after the event date.Figure 1 illustrates the timeline of our event study.

Estimation model(s)
For our main analysis, we calculate AR and CAR of a security using Market Model (MM).We also employ Market Adjusted Model (MAM) and Carhart Four-Factor Model (CFFM) for our robustness checks.See details of formulas and procedures in Online Appendix C.7

Data
We use S&P/ASX 2008 as proxy for market return R m;t under MM and MAM.We collect stock price data of publicly traded companies in Australia whose businesses fall within the scope of the Code and addressing the criteria of a registered news business of the Code (See details in Online Appendix B).
We also cluster the firms into two groups: largefirm group and the group of small-and mediumsized companies (SMEs).Such classification of firms is based on the number of employees and annual turnover by Australian Bureau of Statistics (ABS) and the Australian Taxation Office (ATO). 9he data on the number of employees and annual turnover are recorded based on each firm's annual report of 2019, the year preceding the year the event study is conducted.
As a result, we obtain a sample of total ten news media companies that address the aforementioned criteria as shown in Table 2, of which six are largesized companies while the other four are SMEs.Their daily stock price data during the event window are collected from WRDS's Compustat Global.
In addition, as we complement our studies by assessment of market capitalization of CAR, we retrieve the data on the number of outstanding shares for each company from Thomson Reuters Eikon.Table 3 and Figure 2 illustrate the yearly market capitalization10 of the selected news media companies for this study from 2019 to 2022.
Finally, following Costa et al. (2014), we use Asian-Pacific region (except Japan) risk factors data retrieved from Kenneth R. French data library to estimate CFFM for our robustness check.

The Code and firm performance
What is the channel through which the impact of the Code on the news media stock market is materialized?Recall from Section 3.1.1 the event window spans over 10 months from drafting to passage and includes a series of consultations with businesses and relevant stakeholders.Through such consultations, investors in the news media industry are informed of the upcoming legal provisions that regulate the companies they invest in.How such investors perceive the impact of the Code affects the rise and fall of share price, which in turn affects a company's market capitalization and therefore its market value.A positive reaction of the stock market likely indicates higher investor confidence in the financial performance of the news media industry.The Code, therefore, sends positive signals to the stock market.
The impact of the Code on firm performance is documented through CAR results in percent value (%) and in equity value (AU$) in Tables 4, 5 , 6 and 7 over the three event periods around  20 April 2020, 9 December 2020, 23 February 2021, and the full event window from 20 April 2020 to 23 February 2021, respectively.We find statistically significant and positive CAR for the full event window from the date of the Australian Government's mandate to develop a mandatory Code until its official passage, across all news media firms collectively as a group, and across clusters of firms by size, as shown in Table 7.This implies the collective impact of the three associated event periods as identified within the full event window on the news media sector in Australia is positive and statistically significant.Overall, a compensation scheme put in place, through the introduction of the Code, for core news produced by news media firms at the expense of digital platforms, is perceived as a positive news by investors in the stock market.However, the reactions of investors across each event period, each sub-event window and for individual news media companies do not consistently follow the same pattern.3,537,750 This table documents CAR in percent value (%) and in equity value (AU$) for individual news media firms and groups of firms by size, during the first event period around 20 April 2020, the day when the Australian Government directed the ACCC to develop a mandatory Code to address the bargaining imbalances between digital platforms and news media businesses in Australia.CAR is recorded for three sub-event windows: T0 -on the event date of 20 April 2020; three-day sub-event window T1[−1,+1] including one trading day before and one trading day after the event date and five-day sub-event window T2[−2,+2] including two trading days before and two trading days after the event date.CAR is calculated across all sample news media firms (Group (All)), across largesized firms (Group (Large)), and small-and medium-sized firms (Group (SME)) by taking the average of CAR of the corresponding groups.T-statistic and p-value are calculated to specify the statistical significance of CAR for individual companies and for groups under each sub-event window at 5% (**) and 10% (*) level.
The first event period around 20 April 2020 marks the time when the Australian Government directed the ACCC to develop a mandatory Code to address bargaining imbalances between digital platforms and news media businesses in Australia.The intent of the Australian Government, before this date, had always been a voluntary mechanism for commercial deal negotiations between digital platforms and news media businesses.In April 2020, the ACCC expressed their pessimism over the likelihood of any voluntary agreement could be met by businesses, while the pressure faced by the Australian media sector was exacerbated against the backdrop of COVID-19 pandemic.Therefore, a mandatory Code, instead of a voluntary Code, was expected to incentivize the materialization of actual commercial deals.This news likely surprised stakeholders, including investors of news media businesses, who possibly perceived this as a positive news based on their expectations of future cash flows from potential commercial deals into news media firms.
In fact, we do find positive and statistically significant CAR on the event date (T0) across all companies as a group, and for each group of firms clustered by size, as shown in Table 4.We This table documents CAR in percent value (%) and in equity value (AU$) for individual news media firms and groups of firms by size, during the third event period around 23 February 2021, when the Australian Government announced final amendments to the draft Code prior to its official passage.CAR is recorded for three sub-event windows: T0 -on the event date of 23 February 2021; three-day sub-event window T1[−1,+1] including one trading day before and one trading day after the event date and five-day sub-event window T2[−2,+2] including two trading days before and two trading days after the event date.CAR is calculated across all sample news media firms (Group (All)), across large-sized firms (Group (Large)), and small-and medium-sized firms (Group (SME)) by taking the average of CAR of the corresponding groups.T-statistic and p-value are calculated to specify the statistical significance of CAR for individual companies and for groups under each sub-event window at 5% (**) and 10% (*) level.
estimate CAR for three sub-event windows: T0on the event date of 20 April 2020, T1[−1,+1] -one trading day before and one trading day after the event date and T2[−2,+2] -two trading days before and two trading days after the event date.Interestingly, the same pattern of the empirical results holds true for all three sub-event windows T0, T1 and T2 under the first event period for the group of small-and medium-sized companies, while CAR of large-sized companies is negative during T1 and T2.A possible reason for this may be partly attributed to the existing financial situation of the large-firm group.Of the six large-sized companies in the sample, we drop Southern Cross Media and OOH Media since both companies raised capital around event period 1, which generally decreases stock prices, and hence biases results for our analyses.Of the rest four large companies for our test in event period 1, HT&E and Seven West Media claimed huge losses in the first half of 202011 , which is partly reflected in their negative CAR, albeit statistically insignificant in T1 and T2.The decisive move of the Australian Government seems to have sent a strong signal to the stock market on the exact event date T0 when CAR is recorded positive for all individual firms (with the exception of HT&E), but could not possibly outperform the adverse impact of COVID-19 during sub-event windows T1 and T2.
The second event period is around 9 December 2020, when the draft Code was introduced to the Australian Parliament.This is a significant milestone in the law-making process in Australia, after an extensive period of pre-Parliament consultation process, marking one step closer to the official passage of the Code.Compared to the initial version, this draft reflected several changes in response to stakeholders' submissions.In particular, key concessions were made following submissions by digital platforms, including the acknowledgement of the value of traffic referred by the digital players to news media website from sharing their content during arbitration, and adjustments to notifications around advanced warning of algorithm changes and data.On the one hand, investors in news media companies might perceive the introduction of the draft Code to the Parliament as a positive news, as there is more likelihood that cash flows expected from the remuneration for news contents from digital platforms are going to be materialized.On the other hand, the uncertainty on the implications of the Code's amendments and concessions makes it hard to predict the reaction of the investors in the news media companies.As a result, we find mixed empirical results in the second event period as shown in Table 5. CAR is found positive and statistically significant for the group of all news media This table documents CAR in percent value (%) and in equity value (AU$) for individual news media firms and groups of firms by size, for the whole regulatory event window from 20 April 2020 until 23 February 2021.CAR is recorded for three sub-event windows: T0 -the sum of CAR on the three event dates of 20 April 2020, 9 December 2020 and 23 February 2021; T1[−1,+1] -the sum of CAR in T1 of the three event periods; and T2[−2,+2] -the sum of CAR in T2 of the three event periods.CAR is calculated across all sample news media firms (Group (All)), across large-sized firms (Group (Large)), and small-and mediumsized firms (Group (SME))by taking the average of CAR of the corresponding groups.T-statistic and p-value are calculated to specify the statistical significance of CAR for individual companies and for groups under each sub-event window at 5% (**) and 10% (*) level.
companies and large-sized companies on the event date of 9 December 2020 (T0), while such results for SME group and for all groups over three subevent windows T0, T1, and T2 are negative.One reason for a more positive reaction of large-sized firms than SMEs on the event date of the news may be attributed to the fact that large firms are likely stronger advocates, backed by a stronger corporate legal team during the pre-Parliament consultation process, in comparison with the modest advocacy capacity of SMEs, who were likely fully occupied with survival during COVID-19 pandemic.
The third event period is around 23 February 2021, when the Australian Government made a public announcement on the final amendments to the draft Code, 2 days prior to the official passage of the Code and after a long debate the week before between the Government and digital platforms.Digital platforms are given greater control over the negotiation process, in that they may not be subject to mandatory steps set out by the Code if they are deemed by the Government to have struck sufficient deals with news media companies.As the Code is unclear on how many deals are deemed sufficient, the practical designation of online platforms subject to the mandatory implementation of the Code and the designation timeframe are also unclear.
This uncertainty about the implications and implementation of the Code makes it hard for investors to assess cash flows.In addition, several deals had already been underway since the beginning of February 2021 at the backdrop of the Code being passed. 12While it is plausible to expect a positive reaction from the stock market to the Code being finalized and taking effect, a negative reaction is also possible as some investors, especially those investing in large companies that had secured a deal with digital platforms, may not see much of an impact from the Code, at least for the upcoming short period of time when the existing remunerations are not renewed.The CAR results found on T0 as shown in Table 6 support the latter interpretation, in that the results are negative for the group of all companies and large-sized companies, while positive for SME group, albeit statistically insignificant.However, CAR results in T1 and T2 for all groups are found statistically significant and positive, which likely implies the prevailing effect of the overall positive expectations for the Code from investors upon its passage.

The code and firm size
Another area that this study looks into is the inclusivity of the Code, in that whether news media companies regardless of large or SMEs equally benefit from the Code.Recall from section 2.3 where we discuss the extant state of Australia's media industry as one of the most highly concentrated in the world.In such a context, there is concern that larger companies, which potentially possess higher corporate bargaining power (e.g. in terms of revenues, profits, resources for content creation, market campaign, policy advocacy, etc.) in comparison with smaller companies, are in a more advantageous position to strike more sizable deals with digital platforms (PIJI 2022).While the Code is intended to address the competition imbalance between digital platforms and news media industry, there is an uneven playing field within news media industry itself: news businesses with increased financial capacity from more deals, and those with less or even without any deals, leaving the latter at a distinct disadvantage.Such concerns were exacerbated during the development and passage time of the Code, as only a few large-sized firms had reached some deals with digital platforms by the time the Code was passed in February 2021. 13 In stock markets, how investors absorb information affects the rise and fall of share price, which in turn affects a company's market capitalization and therefore its market value.A larger company has either higher share price or larger number of outstanding shares or both, in comparison with businesses of smaller size.As market capitalization, by definition, is the multiplication between share price and number of shares, 1-% unit of CAR of a larger company is larger in absolute term, compared to that of a smaller 12 For example, Google, through its News Showcase program launched in Australia on 5 February 2021, reached commercial agreements with News Corp., Nine Entertainment and OOH Media's Junkee Media on 17 February 2021, while Seven West Media signed letters of intent with Google and Meta on 15 February and 23 February respectively, of which the duration of each deal is between 3 and 5 years. 13See footnote 12 as above.
company.Taking into account the high concentration of media ownership in Australia, the higher market value of larger companies as perceived by the market therefore likely reinforces their existing market position.As part of this study, we empirically disentangle the impact of the Code on the financial performance of two groups of news media companies in Australia clustered by size: large-firm group and SMEs, by assessing the direction of CAR results on each group, and its magnitude in equity value.
In terms of the impact direction, both groups share the same positive sign during the full event window from the Government's mandate to create the Code in April 2020 to its official passage in February 2021, and consistently over T0, T1 and T2.This means, investors in news media firms, regardless of firm size, generally welcome the introduction of the Code.However, individual event periods have slightly different impacts on each group of firms, particularly in the first sub-event window T0 under event period 2 and 3.In event period 2 as shown in Table 5, when the draft Code was introduced to the Australian Parliament, CAR is found positive and statistically significant for the group of large-sized companies on the event date of 9 December 2020 (T0), while such results for SME group are negative.One possible reason for a more positive reaction of large-sized firms than SMEs', may be attributed to the fact that large firms are likely stronger advocates, backed by a stronger corporate legal team during the pre-Parliament consultation process, in comparison with the modest advocacy capacity of SMEs, who were likely fully occupied with survival during the COVID-19 pandemic.In event period 3 as shown in Table 6, when the Australian Government made a public announcement on the final amendments to the draft Code on 23 February 2021, we find negative CAR for large-firm group and positive CAR for SME group in T0.Most large companies had reached some agreements on a compensation deal with Google and Meta by or before this date.Therefore, it is possible that investors in such companies may not see much of an impact from the Code, at least for the short period of time when the existing remunerations are not renewed, while SMEs may be more eager to see the Code being finalized and taking effect.
In terms of the magnitude of the impact, as proxied by the market capitalization of CAR, there is a gap between large companies and SMEs.As shown in Table 7, during the full event window, we find the market value gain of large-sized group is 10 times, 44 times and 4 times higher than that of SME group over three sub-event windows T0, T1 and T2, respectively.Such gap is at its peak during sub-event window T1 and T2 in event period 3, when the market value gain of large-sized group is 74 and 49 times respectively higher than that of SME group, as shown in Table 6.Interestingly, even though SME group records higher CAR than large-firm group in percent value (%) during the full event window, as shown in Table 7, their gain in equity value (AU$) is much smaller than that of larger firms.Such observation is expected as it reflects the existing high concentration of media ownership in Australia.
The Australian Government might have foreseen the vulnerability of SMEs in the news media sector once the Code was introduced.During implementation of the Code, the ACCC grants authorization for collective bargaining between relevant media associations, whose members are mostly SMEs, and the digital platforms, to help empower the negotiation capacity of those smaller players, while avoiding potential breaching of competition law during payment negotiation process.For example, Country Press Australia (81 members and 161 regional newspapers) or Commercial Radio Australia (261 members).A lesson learnt for governments who are considering the Australian approach is to factor their current media landscape in policy decisions.If there is no one-size-fit-all policy, it is recommended to have a set of policies and supporting regulatory tools, to ensure achievement of key policy objectives, with minimal unintended impacts on vulnerable groups in the society.

V. Robustness
We perform several robustness checks.First, we estimate CAR using MAM and CFFM following Equation 7 and 8 in Online Appendix C 14 , as alternative approaches to MM used for our main analysis.We compare CAR obtained from MM, MAM, and CFFM for the full event window, as well as for each sub-event window T0, T1 and T2, as shown in Table 8 in Online Appendix D. 15 We find a generally similar pattern of CAR obtained from these models, albeit slight different in the magnitude.This indicates the robustness in our empirical findings, which are not an artefact of a particular estimation model.
We also perform a placebo test to ensure our main results are not driven by unobserved factors that happen simultaneously with the Code.Following Barrett and Chen (2021), we construct a placebo sample of companies, which i) are publicly traded on ASX; ii) have similar size in terms of revenue in 2019 (the year before the identified event window); iii) are of other business sectors than news media industry and iv) have no corporate or sector-specific events that may impact the stock returns of those companies around the event periods used in our main analysis.We repeat our event-study estimation on this placebo sample using the same event window, expecting a null result: there should not be a statistically significant change in stock index returns in those companies outside media industry which are not affected by the Code around the dates associated with the Code.Our findings are in line with our expectation, as shown in Table 9 under Online Appendix D. 16

VI. Conclusion
We examine the changes in the financial performance of Australian news media firms in the stock market following the introduction of the Code using Event Study Methodology.We find a collectively positive reaction of the news media industry to the introduction of the Code, although the reactions are slightly different amongst individual firms and groups by size to the different stages of the Code's development.We disentangle the impact of the Code on two groups of firms by size: large firms and SMEs.We find investors in both groups welcome the introduction of the Code.Our robustness tests reinforce our empirical findings.
In such a global context where the dominant position of giant online platforms vis-à-vis the news media industry is under scrutiny by governments around the world, our findings inform policymakers about an avenue for regulatory reform in order to sustain the legitimate interest of the news media businesses and consumers in digital markets.Furthermore, when considering a similar approach to regulate Big Tech companies, governments should take into account the current state of the country's media landscape, to ensure the inclusivity objective of any policy choices.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Timeline of regulatory event study.

Figure 2 .
Figure 2. Market capitalization of selected news media companies, 2019 -2022 (in thousand AU$).This figure depicts the yearly market capitalization of selected news media companies from 2019 to 2022, calculated by multiplying the yearly price of a stock by its total number of outstanding shares.

Table 1 .
Key event dates associated with the development of the Code.

Table 2 .
Selected news media companies for empirical analysis.

Table 4 .
CAR and market capitalization of CAR, Event Period 1-20 April 2020.

Table 5 .
CAR and market capitalization of CAR, Event Period 2-9 December 2020.

Table 6 .
CAR and market capitalization of CAR, Event Period 3-23 February 2021.

Table 7 .
CAR and market capitalization of CAR, Full regulatory event window.