Ex Ante: The Case of Japan

Introduction

The rise of the digital economy has reopened debate on the most fundamental assumptions of competition law. Traditional competition enforcement rests on identifying market power, examining anti-competitive conduct, and intervening only after an infringement has occurred. This structure suited the industrial-age conception of competition that turned on price and quantity. Yet the economic power, data dominance, and network effects attained by digital platforms now call into question whether this approach remains sufficient in certain markets.

Companies such as Apple, Google, Amazon, Microsoft, and Meta are today no longer merely undertakings active in specific markets. They have become the gatekeepers of digital ecosystems through which millions of businesses reach their customers, applications are distributed, payments are processed, data is handled, and artificial-intelligence systems are developed. Whether a developer can reach users, or whether a start-up can scale, often depends on the rules these ecosystems impose.

In the face of this transformation, many countries are developing new regulatory models. The European Union’s Digital Markets Act (DMA) is the best-known example; however, Japan too has adopted a distinctive approach of its own in competition law. The model Japan has chosen is frequently described in the literature as an “ex ante digital competition regime,” and because it is an implemented, current application it can be observed not only theoretically but also empirically.

This study aims to assess the basic structure of Japanese competition law, how the ex ante approach is applied to digital platforms and the artificial-intelligence ecosystem, the concrete obligations it imposes, the actual responses of Apple and Google to those obligations, and the criticisms directed at the model.

Foundations of Japanese competition law

The foundation of Japanese competition law is the Antimonopoly Act (AMA) of 1947. The authority responsible for its enforcement is the Japan Fair Trade Commission (JFTC). The JFTC operates in areas such as combating cartels, resale price maintenance, abuse of a dominant position, the review of mergers and acquisitions, and the prevention of conduct that may lead to monopolisation.

For many years Japanese competition law was administered in a manner similar to the classic ex post practices of the European Union and the United States. Under this model the competition authority investigates an alleged infringement, defines the relevant market, analyses the anti-competitive effects, and only then imposes a sanction. Because the method requires concrete proof of an existing harm, it is robust in terms of legal certainty but fragile in terms of speed.

This fragility has become more visible in digital markets. Within the AMA framework, formal sanctions in non-cartel areas — particularly in cases of exclusionary monopolisation and abuse of dominance — have been observed to be exceedingly rare. Indeed, in the period after 2010 a fine for exclusionary private monopolisation could be imposed in only a single case.

The most concrete indicator of this picture is that the JFTC’s first formal sanction against a major technology company came only in April 2025, against Google. That decision targeted MADA- and RSA-type agreements Google had concluded with device manufacturers; however, it was based not on abuse of dominance (monopolisation) but on the relatively easier-to-prove unfair trade practice (UTP — “transactions on restrictive terms”) provision. Moreover, this first sanction did not even include a fine.

Why the JFTC preferred the UTP provision over monopolisation is telling of the limits of classic tools in digital markets: the high evidentiary threshold imposed by market definition and anti-competitive effects analysis pushed the authority toward a more flexible footing. It is precisely from this problem of “late and difficult” intervention that the need for an ex ante regime arose.

The new problem created by digital platforms

In digital markets competition often takes place not through price but through data, algorithms, user base, network effects, and ecosystem control. The more a user is tied to a platform, the more valuable that platform becomes; this in turn creates a barrier that is difficult for new entrants to overcome. Companies that own an app store, an operating system, or an AI infrastructure can, over time, make their rivals’ market entry effectively harder.

The essence of the problem is this: whether an application can reach users, whether a payment system can be chosen, or whether an AI application can be distributed often depends on the extent to which a handful of large platforms permit it. These platforms are actors that simultaneously operate the market’s infrastructure and compete in that market with their own products. This “referee and player as the same party” situation gives rise to a structural conflict of interest that traditional competition analysis struggles to capture case by case.

Japan’s core finding is that, in such markets, waiting for an infringement to occur often amounts to a belated intervention from the standpoint of protecting competition. Once network effects have consolidated and rivals have withdrawn from the market, a sanction imposed afterwards is usually insufficient to restore competition. For this reason, the 2023 Market Study Report on Mobile Operating Systems and App Distribution explicitly emphasised that the AMA alone might be inadequate to ensure timely intervention, and laid the groundwork for ex ante regulation.

The ex ante approach: concept and logic

The Latin-rooted term ex ante means “beforehand.” Classic competition law is ex post, that is, it intervenes after an infringement. In the ex ante model, by contrast, the regulatory authority designates certain digital platforms in advance and imposes directly applicable (self-executing) conduct obligations on them. These obligations take effect without the concrete effect of an infringement having to be separately proven.

The most practical way to understand this approach is through an analogy: rather than waiting for the accident to happen and then punishing the responsible party, the competition authority sets the traffic rules in advance. The aim is less to punish than to prevent anti-competitive outcomes from arising in the first place. This logic offers two important technical conveniences: first, the authority is freed from having to redefine the relevant market in every case; second, instead of a detailed analysis of anti-competitive effect, the question becomes whether the conduct falls within a prohibited category.

The cost of this convenience is a loss of flexibility. When form-based (per se-like) prohibitions are adopted instead of effects-based (rule of reason) analysis, conduct that in a concrete case actually enhances competition may also fall within the scope of the ban. The Japanese model does not eliminate this tension entirely; however, unlike the DMA, it seeks a middle path by allowing designated operators to escape certain obligations by demonstrating a “reasonable justification” or “legitimate reason.”

The Mobile Software Competition Act (MSCA)

The most concrete example of the Japanese ex ante model is the Mobile Software Competition Act (MSCA). The official name of the statute is the “Act on Promotion of Competition for Specified Smartphone Software” (Act No. 58/2024); its informal name is the “Smartphone Act” or, by its Japanese abbreviation, “Sumaho Shin-hō.” The Act was passed in the Diet on 12 June 2024, and its prohibitions and obligations entered fully into force on 18 December 2025.

The MSCA is frequently referred to as “Japan’s DMA” and holds particular significance as the first ex ante digital competition regulation in Asia to be enacted and actually enter into force. Similar initiatives came onto the agenda in India and South Korea but largely slowed down, in significant part owing to trade negotiations conducted with the United States; Japan, by contrast, brought the regulation to life with determination.

The Act targets four types of software regarded as essential for the use of smartphones: operating systems, app stores, browsers, and search engines. Undertakings that provide this software and have reached a scale capable of excluding or controlling the activities of rivals are designated, according to criteria such as the number of users, as “designated operators” (Specified Software Operators). The scope of the regulation is deliberately narrow; the aim is to target not the whole sector but only structural gatekeepers.

The obligations imposed, and a comparison with the former position

To understand the significance of the MSCA it is helpful to compare the obligations it imposes with the practice prevailing before December 2025. Before 2024, the mobile ecosystem in Japan was tightly managed: third-party app stores were not permitted on iOS, apps selling digital content were compelled to use Apple’s billing system, and developers could not steer users to pay over the web. Consequently, neither a price comparison nor an external payment link was possible within an app.

In the field of alternative app stores, the MSCA prohibits platforms from forcing users to use only their own stores. This seeks to prevent app distribution from being controlled through a single channel. With the entry into force, Apple opened the door for the first time in iOS history to third-party app marketplaces in Japan; on the Android side, the Google model — already relatively open — continued to permit third-party stores.

In terms of alternative payment systems, developers are no longer to be compelled to use only the platform owner’s payment infrastructure. Developers may now offer their own payment systems within the app or direct the user to a website to complete the transaction. In practice, Apple has required that alternative payment options always be presented alongside its own In-App Purchase method, so that it remains transparent when the user is transacting through Apple.

Within the framework of anti-steering rules, platforms can no longer prevent users from being directed to alternative payment channels offering more advantageous terms. Providing external links and making price comparisons has become permissible. Under the self-preferencing prohibition, platform operators may not position their own products and services advantageously relative to rivals; this prohibition is especially critical with respect to algorithmic ranking systems.

In addition, user choice screens have been introduced: both Apple and Google must present the user with browser and search-engine options upon setting up a new device or updating the operating system. At the operating-system level, third parties are expected to be granted access to essential hardware and software features such as NFC and biometrics.

The data-use prohibition (Article 5)

The most concrete form of the self-preferencing prohibition is seen in the data dimension. Article 5 of the MSCA prohibits designated operators from using data obtained through software to advance their own products or services in competition. The July 2025 Guidelines illustrated the types of data falling within the scope of this prohibition: user characteristics (place of residence, name, age, gender), device identifiers (account ID, IP address, advertising ID), and information required for payment (credit card number, payment-service-provider account numbers).

The prohibition has an important limit: data that is generated independently of such use and is publicly available — that is, data a business user can already access via a web page, an app store, or a market-information service — does not fall within this scope. The JFTC, acknowledging that external auditing of data use is technically difficult, expects operators to establish effective internal control systems to safeguard compliance and regards such systems as “crucial.”

The sanctions regime and the compliance mechanism

On paper, the MSCA’s sanctions regime is highly deterrent. Where the Act’s core prohibitions are infringed, the JFTC may impose an administrative fine (surcharge). The fine is calculated by a fixed formula: twenty per cent of the turnover derived from the goods or services subject to the infringement. This rate is far above the AMA’s traditional unfair-trade-practice fines (one to three per cent for wholesale and retail) and gives concrete form to the ex ante regime’s claim to deterrence.

An example helps to grasp the magnitude of this rate: had Apple been found, without a valid justification, to prevent other app stores from operating on iOS, it could have faced a fine of twenty per cent of the turnover derived from its app-store business in Japan. For an infringement lasting one year, this figure is said to be capable of reaching the level of JPY 100 billion (approximately EUR 540 million). Alongside the fine, the JFTC may issue cease-and-desist orders, and the way is open for private damages actions and interim-relief requests.

That said, compliance is secured primarily through annual reports and ongoing dialogue. Designated operators must submit a report each year detailing the compliance measures they have taken. The JFTC has positioned this model as a dialogue-based supervision that, rather than sudden and severe sanctions, prioritises securing the correction of conduct by engaging with the operator when a risk of infringement emerges.

The actual responses of Apple and Google

What lifts the MSCA out of being merely a text is that it has been implemented and is observable. Just before the entry into force on 18 December 2025, both companies announced their compliance steps. Apple delivered the changes as part of the iOS 26.2 update in a Japan-specific manner, introducing alternative app marketplaces, payment processing outside the App Store, and a browser choice screen. Because Android is already relatively open, Google’s changes remained more limited; browser and search-engine choice screens, expanded billing options, and the side-by-side comparison of external and internal payment methods came to the fore.

Apps distributed outside the App Store will not undergo a full App Review process; instead they will be subject to a “notarization” process focused on basic functionality and on known malware and serious security threats. Apple has emphasised that alternative payment options will always be offered alongside its own In-App Purchase method, and that protections such as refund support and subscription management will apply only to transactions made through Apple.

The fee structure has been the most contentious heading in implementation. Apple has reduced the App Store commission in Japan to ten per cent for small developers and subscriptions, and has announced a new structure in which total fees vary roughly between five and twenty-six per cent depending on the distribution and payment method. In the past the standard commission was thirty per cent, and fifteen per cent for small developers. Google, for its part, has retained the thirty per cent commission for in-app purchases made through its own payment system; for external payment steering, which it had previously banned outright, it has introduced a commission of twenty per cent.

On 17 February 2026 the JFTC published the compliance reports of Apple and Google. The reports set out concrete openings such as third-party app marketplaces accepted for the first time, external payment links, browser and search-engine choice screens, and data-portability mechanisms. However, the JFTC requested further explanation regarding fee levels and the basis on which those fees are calculated; this meant the reports received not an “unconditional pass” but a mixed response.

Criticisms and debates

The MSCA is not a regulation adopted with a single voice; it is the subject of serious debate from both legal and economic standpoints.

First is the criticism of the form-based (per se) approach. Some commentators argue that the MSCA’s adoption of categorical prohibitions — departing from the AMA’s case-by-case, effects-based analysis — disregards the circumstances of the concrete case and may deter innovation. On this view, competition already exists between Apple’s integrated ecosystem and Google’s open Android model, and generative AI together with new devices is invigorating that competition further; hence the assumption of an “entrenched monopoly” is debatable.

Second is the security tension. Third-party stores, unvetted applications, and payment links opening to the outside may give rise to risks of phishing, fraud, and data security. It is therefore argued that operators must be able to set reasonable limits and warn users, and that the regulation should be carried out consistently with Japan’s 2022 National Security Strategy. Apple’s emphasis on these new risks in its statements is a reflection of this tension.

Third is the criticism of de facto nullification. Although developers have gained new options such as external payment and alternative stores, the commissions the platforms have attached to these channels (including rates of up to twenty per cent for external payment) limit the overall reduction in cost. When third-party payment-processing costs are added, it is argued that, for many developers, departing from standard billing is not economically meaningful; hence the door the law opened remains narrow in practice.

Fourth is the concern about under-enforcement. It is noted that the JFTC has historically imposed very few formal sanctions on major technology companies, and that South Korea’s competition authority (the KFTC), of comparable institutional capacity, pursued far more infringement cases against digital platforms over the same period. The fact that the KFTC has been able to impose high fines in technically complex files such as Qualcomm shows that smaller authorities can also act ambitiously, reinforcing this criticism directed at the JFTC.

Fifth is the geopolitical dimension. In documents published at the U.S.–Japan summit, the U.S. administration requested that the law be applied in a manner that does not create discrimination against U.S. companies and that preserves the balance between user security and fair competition. This shows that digital-competition regulation has now become not merely a legal matter but also a part of trade diplomacy.

The Japanese model from the standpoint of artificial intelligence

One of the most forward-looking aspects of the Japanese regulation is that it considers not only existing mobile platforms but also the artificial-intelligence ecosystem of the future. Competition in the field of AI today is shaped through data sets, processing power, cloud infrastructures, app stores, and user access. Whether an AI application can reach millions of users often depends on the infrastructures and distribution channels controlled by large technology companies.

For this reason, a significant part of the debate conducted in Japan is concentrated around the following questions.

• Can access to AI models and developer interfaces be blocked?
• Can a data advantage become decisive to a degree that structurally distorts competition?
• Can a platform, through ranking and default settings, favour its own AI products?
• Can app stores make the market entry of AI ventures more difficult?

Conclusion

The ex ante digital competition regime developed by Japan shows that the boundaries between classic competition law and sectoral regulation are increasingly blurring. In this model the aim is not to punish, after the fact, the structural market power held by digital platforms, but to prevent anti-competitive outcomes from arising in the first place. The MSCA has given concrete form to this philosophy through designated operators, directly applicable prohibitions, and high-rate deterrent fines.

At the same time, the regulation’s experience in its first months also reveals that the ex ante model is not a magic wand on its own. The ongoing debate over commission rates shows that there may be a serious gap between the clarity of the rules on paper and their actual effect in the market. In other words, the success of an ex ante regime depends not only on the existence of rules but on the extent to which those rules align with economic reality and on the authority’s determination in enforcement.

Especially today, as artificial-intelligence technologies develop rapidly, it is clear that control over data and infrastructure will move to the centre of competition law. For this reason the Japanese example is a living laboratory that must be followed closely not only with respect to mobile app stores but also from the standpoint of artificial intelligence, the data economy, and the future of digital platforms.

Can competition law still make do with ex post intervention to keep pace with the speed of the digital economy, or will the law of the future inevitably evolve toward an ex ante structure?