Lately, there seems to be a race among competition and regulatory authorities to see who can come up with the most far-reaching proposal for regulating platforms. The seemingly endless saga of access obligations and price controls has made the European digital market extraordinarily messy and the region’s digital policymaking into something close to a zero-sum game. Each new intervention must go further than the last. Actions accumulate up to something so complex and interventionist that it is difficult to see how an authority can make sense of it all. The platform conduct requirements by the UK Competition and Markets Authority (CMA) are a case in point.
Just before summer, the CMA consulted on proposed steering conduct requirements (CRs) for Apple and Google. As we argued at the time, those proposals imported concepts from utility and access-price regulation into app-store ecosystems and used them to determine the platforms’ cost structures. The CMA’s chosen comparators were legacy infrastructure sectors built around stable, identifiable assets and long histories of state ownership. None of them resembles an integrated and rapidly evolving mobile platform, which is what made the remedy odd in the first place.
The NFC Access CR: Another Design-Access Intervention
The CMA is now back with a similar proposal. On October 20, 2026, it published a consultation on a CR that would require Apple to open access to near-field communication (NFC) on iOS devices. The stated objective is to ensure effective, interoperable access to NFC on ‘fair, reasonable, transparent and non-discriminatory’ terms, so that developers can improve their existing offerings or develop new uses of NFC technology (CMA 2026, paragraph 2.2). In the abstract, that objective is well understood. The difficulty lies in its means. The CMA’s CR would intervene directly in how access is provided.
Apple would have to offer developers a choice between Host Card Emulation (HCE) and Secure Element (SE), extend access to features such as Express Mode and Power Reserve, and justify its access fees against detailed pricing principles to the CMA’s satisfaction. The proposal would also regulate authorisation procedures and require an interface allowing developers to check their default contactless app status without restrictions on the frequency of requests (Section 3, paragraphs 15-19 and 24-26, p. 28-30). Through these changes, it is evident that CMA will gain considerable oversight of Apple’s technical, commercial and administrative arrangements for providing NFC access, extending its influence into the wider architecture of mobile payments on iOS.
Section 20 of the Digital Markets, Competition and Consumers Act 2024 (DMCCA) allows the CMA to impose requirements concerning fair trading, interoperability and preferential treatment. However, the proposed NFC intervention raises an important scoping problem. While the CMA designated Apple’s mobile operating system as a Strategic Market Status (SMS) activity, it expressly excluded Apple Pay and Apple Wallet (paragraph 1.20, footnote 25). Yet the commercial target of the NFC Access CR is precisely that wallet.
This point of designation should prompt a cautious and proportional treatment by the CMA. Much of its evidence and arguments are about the digital payments market, not operating systems. The CMA, for instance, expects the new entry to place downward pressure on the per-tap fees that card issuers pay Apple for Apple Pay transactions, which it put at up to £150 million for 2025 (paragraph 5.90(b)). It describes high NFC access fees as a means of insulating Apple’s own fees from competitive pressure (paragraph 4.129(b)). Regardless, the merits of these points, these points are solely outside the scope of the iOS.
Combining designation complexities with radical conduct requirements raises problems of legitimacy and concerns of overreach. This mix risks turning access regulation into a tool for making competing business models commercially viable, rather than addressing demonstrably unfair barriers to competition. The designation of the OS is being used as a lever to influence a downstream service that the CMA itself left outside the designation.
The proposal also draws heavily on what the European Commission did with Apple in 2024. In July of that year, the Commission accepted legally binding commitments from Apple to close an investigation about the ability of competing wallets to access NFC on iOS. The Commission’s concerns were twofold: that Apple was vertically foreclosing competing in-store mobile wallets in order to protect the transaction fees it earns from Apple Pay, and that restricted access could foreclose the market by preventing follow-on innovation. To resolve these concerns, Apple created the APIs needed to give third-party wallets equivalent access to NFC in HCE mode – a technology that keeps sensitive credentials off the device and sends the phone a limited number of single-use tokens, rather than relying on the phone’s SE (paragraph 1.11(a)).
This required a change in the underlying architecture. Because Apple Pay relies on the hardware-based SE, it had to build and maintain a parallel, software-based route specifically to comply in the European Economic Area (EEA). The European Central Bank (ECB) criticised the exclusion of SE access to third parties, arguing that HCE alone was insufficient for applications such as offline digital euro payments, which require tamper-resistant hardware. In the UK, that concern is already met: Apple has offered SE access to third parties since October 2024 (paragraph 1.13). What the CMA adds now is an obligation to provide HCE as well – the technology that the ECB considered insufficient on its own, and to let developers choose between the two. In this way, the CR goes considerably further.
Since designating Apple with SMS in October 2025, the CMA has become fixated on one observation: that new wallets entered the EEA market after the EU intervention – a regulatory effect – while none has entered in the UK after Apple opened SE access. The CMA’s proposal raises three fundamental concerns:
- The case for intervention rests on weak evidence – the CMA treats the absence of competing wallets as a failure of access conditions, without establishing whether lower fees would actually bring more developers or consumers into the market;
- The remedy goes further than the problem identified – rather than addressing the terms of existing SE access, the CMA would require Apple to support two competing technologies, without clearly demonstrating why both are necessary;
- The costs and risks receive too little attention – the proposal would give the CMA considerable discretion over pricing and technical design, while leaving questions about implementation costs, security and consumer preferences unresolved.
Entry Depends on Demand, Not Only on Access Terms
The CMA’s starting point is that, two years after Apple made its NFC & SE platform available in the UK, there is still no evidence of NFC apps launching on it (paragraph 1.13), whereas several wallets launched in the EEA once free HCE access was granted (paragraph 1.14). From this, the CMA infers that Apple’s current fees are ‘too high to support commercially viable entry’ (paragraph 1.16) and that the lack of entry suggests those fees do not reflect the value developers expect from access (paragraph 4.183). Simply put, the CMA equates limited market entry with a failure of access conditions.
This is not a good way to understand the market. In the first place, market entry crucially depends on demand and expected returns, not only on the terms of access. Moreover, the EEA is a poor guide for a more basic reason: the two markets start from different places. When the Commission intervened, third-party wallets had no access to NFC on the iPhone at all, and HCE was the route through which access was created. As we just pointed out, this access already exists in the UK: Apple has offered its NFC & SE Platform to developers since October 2024. Mandating HCE imports the EEA’s technology into a market that does not have the EEA’s problem.
The legal rules also differ. In the EEA, Apple must provide HCE access for free under its commitments and the Digital Markets Act (paragraph 4.263). The CMA, however, would allow Apple to charge for access. It even asks whether developers would still enter the UK market if they had to pay (question 6.13). If the CMA does not know the answer, it cannot assume that the EEA experience predicts what will happen in the UK.
Nor can the EEA example show what HCE achieves on its own. Developers there entered under a package of measures – free access, default-wallet functionality and binding commitments lasting ten years, so the outcome cannot be attributed to any single element. The CMA points to wallets launching within six months as evidence that HCE can encourage entry (paragraph 4.27). Yet it explicitly says that its UK approach will differ from Europe’s, including on pricing (paragraph 4.264).
The evidence from Europe is also limited. Even with all these measures in place, Apple reported only six competing HCE wallets across the EEA in its written response to the Australian House of Representatives Standing Committee on Economics, following a parliamentary hearing in February 2026. This was some 20 months after the commitments took effect, in a market far larger than the UK. The CMA’s own consultation identifies five such providers in the EEA. It is too early to judge the EU intervention, but equally too early to treat it as a success story. A handful of launches under conditions the UK will not replicate is hardly a basis for predicting substantial competitive benefits from a different regime in a smaller market. Yet the CMA assumes that multiple developers will take up NFC (paragraph 5.95(a)) and concludes that the costs of its intervention will be very low relative to the benefits, despite being unable to quantify either.
It is also worth looking at who actually entered the EEA market: PayPal, Klarna, BBVA, Curve and Vipps MobilePay. They are all established payment providers or banks, not new fintech start-ups. An industry analysis of the US market points in the same direction, identifying large payment firms, card issuers and businesses with existing customer relationships as likely beneficiaries of NFC access.
The same logic applies to Europe: these firms already have the customers, banking partnerships and resources to absorb the costs of launching a competing wallet. The CMA itself describes HCE as the ‘clearest route to faster and lower-cost entry for established payment use cases’ (paragraph 4.67). Opening NFC can encourage some entry, but mainly from established firms in adjacent markets rather than new fintech competitors the CMA hopes to attract (paragraphs 1.5-1.6, and 2.3(a)). In this regard, the CMA confuses the promotion of competition with the promotion of particular competitors.
The CMA’s framing also overlooks where competition in mobile payments already takes place. Apple Pay does not charge merchants directly; instead, Apple recovers its investment through fees paid by card issuers. In its 2026 submission to the Reserve Bank of Australia, Apple notes that more than 125 Australian financial institutions, from small credit unions to the largest banks, use Apple Pay on the same terms. Merchants can also route debit transactions through lower-cost networks, while the Reserve Bank treats wallet fees as an eligible cost within its interchange framework. Apple argues that merchants benefit from higher conversion rates, lower fraud losses and lower payment costs. Although the details of the Australian system differ from those of the UK, the example shows how competition can take place within a wallet rather than only between wallets.
Consumers can already choose between cards from competing banks within Apple Pay and switch between them at the point of sale. The CMA’s answer is to create more wallet apps. Yet a new wallet does not necessarily mean a new payment method. UK Finance reports that 90 per cent of mobile wallet users have a debit card set as their default payment method, while the growth of mobile contactless payments is largely replacing physical contactless card transactions. A PayPal or Klarna wallet on the iPhone could simply offer another way to make the same Visa or Mastercard payment. The CMA expects competing wallets to put downward pressure on the fees banks pay Apple, with any savings potentially passed on to consumers (paragraph 5.90(b)). But lower fees for banks are not the same as lower prices for consumers, who already pay nothing directly to use Apple Pay. The CMA just has not shown why adding another layer of wallet competition would deliver benefits that competition between banks and payment providers does not already provide.
The CMA’s own evidence base is also incomplete. Mastercard and Visa announced services to support developers using Apple’s Secure Element platform only in July and September 2026 (paragraph 1.13, footnotes 10-11). The CMA admits that it does not yet know whether these services will encourage developers to enter (paragraph 4.28) and dismisses other explanations Apple put forward from satisfaction with QR codes to limited demand for third-party wallets (paragraphs 4.135 and 4.146).
Instead of giving the existing route time to work, it wants to mandate HCE access as well. But why should a second technology succeed where the first has yet to be properly tested? Easier access does not create consumer demand or make an unprofitable business model viable – the EEA, where HCE access was free, shows as much. The CMA is proposing another route to market before establishing whether the lack of competing wallets reflects restricted access, weak demand or simply a market that has not yet had time to develop.
Why Prescribe another Technology where one is Already Available
The CMA proposes the mandatory provision of both HCE and SE (paragraph 4.12) and provisionally concludes that requiring both is the only effective option (paragraphs 4.13). In this way it prescribes two technical solutions without establishing that the existing SE route is insufficient. If the CMA’s real concern is that SE access is too expensive or its terms are too restrictive, the obvious remedy is to address those terms directly – not to require Apple to build and support a second technology. Payments rest on trust and security, and the CMA proposes to add a less secure route in order to solve what it presents as a pricing problem. Ironically, the consultation itself contains evidence that SE is the stronger technology.
The CMA accepts that SE is better suited for cases requiring secure on-device storage, continuous offline operations or low-power functionality (paragraphs 4.9(b), 4.53 and 4.59); and that it remains suitable for high-assurance credentials (paragraphs 4.46-4.47). Two banks told the CMA that if offered only one technology, they would choose SE, and it is described as a value-added technology (paragraph 4.26). HCE’s advantage lies in the speed and cost of porting for firms that have already built HCE solutions for Android or EEA (paragraphs 4.16 and 4.23). That is revealing. One of the CMA’s stated aims is to encourage new and innovative uses of NFC (paragraph 2.3(b)), yet it acknowledges that SE is particularly well suited to applications requiring offline functionality and secure credential storage (paragraphs 4.9(b), 4.47 and 4.59).
The CMA’s own consultation question 6.2 suggests uncertainty about whether developers would want both technologies once access is priced. The expected rush to HCE is likely to be illusory. However, it is well established that HCE is less secure than the NFC & SE model, which relies on a dedicated encrypted hardware chip; and that many platforms are moving away from legacy HCE implementations towards the model Apple has developed.
The European Commission’s Qualcomm/NXP merger decision (2018), also records longstanding industry concerns about HCE security. Samsung argued that HCE alone was insufficient for high-security financial transactions without additional protections, such as SE (paragraph 161, p. 41). Infineon expected manufacturers to continue relying on SE, arguing that switching to HCE would mean degrading the security of their solutions (paragraph 165, p. 42). Although these assessments are several years old, they show that major industry participants have long regarded hardware-based security as particularly important for sensitive applications.
Moreover, mandating two technologies can get costly. Apple told the CMA that a dual-technology regime would require it to develop and support multiple onboarding processes, separate terms of access, additional developer-relations support and divergent compliance programmes for HCE, and that developers could game the availability of both routes to pay the lowest price (paragraph 5.24). It also argued that requiring HCE alongside SE would undermine its ability to recoup its investments in the NFC & SE Platforms (paragraphs 4.136 and 5.48). The obligations extend well beyond the access route itself. Apple must give third parties interoperable access to enabling technologies used by Express Mode and Power Reserve, and must allow a third-party default app to complete transactions without authentication or unlocking the device wherever Apple’s own wallet would not require it.
The CMA’s CR is a continuing supervision of its payment architecture. Because first it would have to keep both routes aligned with evolving industry standards, build the necessary functionality into the first iOS release after the CMA’s decision, file compliance reports every three months and demonstrate that any interruption of access was justified. It also cannot price the two routes differently without objective justification (paragraph 4.211).
Apple pointed out that extending Express Mode to third parties could require substantial engineering work on its proprietary framework (paragraph 5.25(b)). The CMA acknowledges these costs but assumes they can be recovered through NFC access fees (paragraph 5.45). Yet the same consultation expects those fees to be substantially lower than Apple’s current charges (paragraph 1.17), without establishing how many developers would pay them. This development is highly implausible. Cost recovery depends on both the fees charged and the number of developers using the platform, neither of which the CMA has demonstrated.
The consumer evidence also counsels caution. Which?’s 2026 survey of more than 4000 UK consumers found that only 7 per cent of iPhone users and 8 per cent among Android wallet users had ever switched digital wallet. Android has long allowed competing NFC wallet applications through HCE – yet the difference in reported switching is just one percentage point. Greater flexibility on Android has still not resulted in a market for competing wallets.
Several other factors matter more than access terms – preinstallation, familiarity and satisfaction. 45 per cent of the users who had not switched said they were happy with their current wallet. This is a comparison the CMA needs to confront: if competing wallets have long been available on Android but few consumers switch between them, why should mandating HCE on iPhones produce the competitive pressure the CMA expects?
This makes the CMA’s proposed default-wallet measures particularly questionable. Rather than addressing why consumers rarely switch, the CMA wants to make it easier for developers to persuade them to do so. Users can already choose a third-party default contactless app on Apple’s existing platform (paragraph 4.85), and Apple does not currently limit how often developers prompt them (paragraph 4.249). The CR would go further. Apple would have to let developers check their default status without any limit on the number of requests and would be barred from limiting how often developers prompt users to switch (Section 3, paragraphs 23-26, p. 30).
By comparison, Apple limits default-browser status checks to four a year to guard against excessive prompting (paragraphs 4.247-4.248). The CMA dismisses the risk of over-prompting on the grounds that developers have incentives not to annoy users (paragraph 5.77). But the low switching rates suggest that many consumers are content with their existing wallets, not simply waiting for more opportunities to change them. Nearly half of wallet users chose their wallet because it was already set up on their device. Making it easier for developers to repeatedly ask these consumers to switch does not address the apparent lack of demand for alternatives. It risks turning a policy intended to increase consumer choice into one that makes consumers defend a choice they have already made.
How Does the CMA Know Fees Are Too High?
The NFC Access CR seeks to regulate the prices Apple charges for access through both SE and HCE (paragraph 4.188), yet it declines to say what those prices should be (paragraph 4.121). To facilitate this version of ‘forced’ switching, the CMA has already concluded that NFC access fees should be ‘low and substantially lower’ than today’s (paragraph 4.187), before establishing an objective basis for calculating appropriate fees (paragraphs 4.115-4.187). Yet its own framework allows higher fees, where justified by costs and value (paragraph 4.186). The CMA cannot claim to assess fees on their merits while effectively ruling out one of the possible outcomes in advance. That rests on four pillars – the views of developers who would pay the fees, the absence of entry, comparisons with Google, and its own break-even analysis for a ‘copycat wallet’.
The CMA’s break-even analysis is particularly revealing. It finds that interchange revenues would cover Apple’s fees only under a narrow set of assumptions, and that developers unable to earn interchange revenue would require even lower fees (paragraph 4.144). In the UK, interchange fees on consumer cards are capped at 0.2 per cent for debit and 0.3 per cent for credit transactions, and those revenues go to card issuers rather than automatically to wallet providers.
A competing wallet relying on a share of those fees has little revenue with which to cover its costs, regardless of Apple’s charges. Reducing NFC access fees cannot solve a business model that generates insufficient revenue in the first place. Yet the CMA’s reasoning increasingly treats the commercial viability of different wallet models as a test of whether Apple’s fees are reasonable. Taken far enough, that logic points towards fees close to zero, particularly for developers with no independent source of transaction revenue. The CMA says it is not requiring free access ‘at this stage’ (paragraph 4.163 and footnote 229), but its reasoning makes it difficult to see where a positive fee would be considered acceptable.
Clearly, this also undermines Apple’s incentive to invest. The CMA wants the fees set at the lowest level at which Apple is ‘fairly remunerated’ (paragraph 4.141), while narrowing what counts towards that remuneration. Costs relating to Apple’s own wallet must be excluded, only value attributable to Apple’s own specific investments are to be counted, and fees are expected to be low because Apple did not invent the underlying technologies. Yet Apple’s SE platform, Express Mode and Power Reserve were built for its own wallet – the investment the formula does not consider.
The CMA argues that Apple’s incentive to invest in NFC does not depend on NFC returns, because better features make iPhones more attractive (paragraph 4.142). But this is a remarkable view that confuses the incentive to improve the iPhone with the incentive to invest in a particular technology. Apple can benefit from better NFC functionality without having a commercial reason to develop and maintain two competing solutions, each of which the regulator requires it to provide on terms it cannot freely determine. The CMA just cannot assume that the benefits to iPhone sales will justify any investment it mandates, regardless of its costs or returns. Nor does it properly account for the burden of complying with its pricing rules, treating that burden as justified by Apple’s market power (paragraphs 5.46 and 5.79(c)). That reasoning risks making the intervention self-justifying: the CMA invokes market power to impose pricing controls, then relies on the same premise to dismiss their costs.
The proposed formulas for calculating fees are also unclear and leave considerable room for subjective judgement. The cost-based principle requires Apple to account for its direct and indirect costs and earn a reasonable return (Section 3, paragraph 11(a), p. 26-27). Yet Apple told the CMA that its accounting systems do not allow these costs to be reliably calculated or allocated (paragraphs 4.152-4.153), and argued that cost-based pricing belongs to utility regulation, not a dynamic platform built around intellectual property (paragraph 4.155). As we argued in relation to the steering CRs, better accounting cannot solve the basic problem of deciding how shared costs should be divided.
The value-based principle is no clearer. It requires Apple to calculate the value of its services while adjusting for its market power and the value exchanged with developers (Section 3, paragraph 11(b), p. 27); paragraph 4.182). But there is no clear way to measure either adjustment, as the CMA itself acknowledges (paragraphs 4.181 and 4.184). The CMA’s comparison with patent licensing on fair, reasonable and non-discriminatory (FRAND) terms do not help either: unlike patent licensing, there is no prior licensing commitment or established set of comparable agreements against which to judge the appropriate fee (Section 4, p. 83 and 86, footnotes 210 and 226). Because CMA declines to specify a fee, Apple has no safe harbour.
The flexibility principle is perhaps the most puzzling of all. It requires Apple to offer different fee structures depending on how developers use NFC, whether their applications are new or established, and how they earn revenue. Developers must also be able to choose between these structures and switch as their circumstances change (Section 3, paragraph 11(c), p. 26; paragraphs 4.201-4.205).
In practice, then, Apple would have to offer a list of tariffs suited to different business models, potentially charging less for new applications or those generating little revenue. But why should the price of accessing Apple’s technology depend on how another company chooses to make money? The CMA’s own account is contradictory: it says tariff options should not be based on developer characteristics (paragraph 4.203), yet the CR requires Apple to consider ‘whether and if so, how revenue is generated per NFC tap’ (Section 3, paragraph 11(c)(iii), p. 27).
This effectively makes the developer’s business model a factor in setting fees, creating scope for inconsistent pricing and disputes over how different developers earn their revenues. The CMA also grounds this principle in a statutory power concerned with presenting choices to users, not setting access prices for businesses (paragraph 2.10(a)(iv)), raising questions about the scope of its intervention.
The CMA also wants Apple to offer different pricing options while prohibiting discrimination between similar developers (paragraph 4.198). It is, generously put, unclear where legitimate price differences would end and discrimination would begin, or how Apple could prevent developers from repeatedly switching to whichever tariff costs them least. Remarkably, the CMA dismisses that concern (paragraph 5.41).
How the Draft CR Treats Security
The proposal risks treating security safeguards principally as restrictions on competition, rather than as essential features of payment systems. Apple will be able to interrupt access only in an instance or credible threat of serious harm resulting from a breach of law or industry standards and only where it can show the CMA that the interruption is strictly necessary and goes no further than necessary. The CMA concludes that both HCE and SE can be used securely for payments (paragraph 4.46) and that residual additional security or privacy risk is likely to be low (paragraph 5.72).
That assessment underweights three fundamental security concerns:
- First, the security threat is growing, and remains concentrated where third-party apps can reach NFC functionality. NFC-based attacks on Android smartphones aimed at stealing money grew by 188 per cent in the first four months of 2026 using two schemes – direct and reverse NFC relay, some of which exploit HCE. In late 2024, fraudsters tested a scheme on users, in which victims were socially engineered into installing an app purporting to be a government, banking or other service, then prompted to hold their card to their phone and enter their PIN for verification; the card data was then replayed through HCE on a fraudster-controlled device. In the Netherlands, a scheme combining AI-generated fake bank calls with a malicious NFC app captured card data. Ghost Tap malware persuades victims to tap their card against their phone and relays the card data via a command-and-control server to a criminal-controlled device, which replicates the card through HCE to complete transactions on illicitly obtained point-of-sale terminals. The CMA’s own footnotes list a series of 2025 and 2026 reports on NFC relay malware (footnote 325, p. 125). While these are Android-specific, and the encryption environment is different from Apple, these attacks are more prevalent in an environment where third party apps can access HCE.
- Second, the CMA relies on safeguards that its own proposal could weaken. It points to Apple’s App Review process as protection against security risks and excessive prompting (paragraphs 5.69 and 5.77(b)), and recognises the role of industry self-regulation (paragraph 5.70). Yet the proposed CR also treats app review procedures as potential unjustified restrictions on access (Section 3, paragraph 5(c) p. 24), limiting Apple’s discretion to reject or restrict developers. Fraudulent developers can hide features that are activated only after an app has passed review. In 2024, Apple rejected more than 43,000 app submissions for hidden or undocumented features, and over 320,000 for copying other apps, spam or misleading users. It also reported preventing more than $2 billion in fraudulent transactions and banning over 1.6 million accounts from transacting again. NFC access creates another opportunity for such abuse. The CMA is willing to rely on safeguards provided by wallet developers and card schemes, but appears less willing to trust those provided by Apple itself.
- Third, expanding access changes who is responsible for security. Apple already allows third-party developers to use its SE in the UK, subject to authorisation, security requirements and independent validation. The CMA now wants Apple to provide HCE access as well, introducing another technical route with different security controls. With third-party wallets, security also depends on the developer’s authentication systems and back-end infrastructure. A transaction can be cryptographically legitimate without revealing whether the person behind it is a fraudster. The National Cyber Security Centre told the CMA that, where cards are stored using HCE, card data must be stored securely, preferably outside the operating system, because an operating-system exploit could allow malware to access it (paragraph 5.58(c)).
Apple argued that developers choosing HCE do not fully bear the security and privacy costs of a less secure solution, leaving users exposed (paragraph 5.54). The CMA’s answer, that wallets have strong incentives to protect customers (paragraph 5.61), does not fully address this problem. A security failure in a third-party wallet could also damage trust in the iPhone itself, even if Apple did not develop or operate that wallet. Apple made a related point to the US House Judiciary Committee in 2020, arguing that uncontrolled third-party NFC access for payments would undermine its integrated security architecture.
The CMA appears to forget that Apple’s architecture is not Android’s. Both have their merits, but they are different. The iPhone is built on an integrated, proprietary model, and consumers choosing an iPhone know they are entering the walled ecosystem. The CMA’s own consumer research found that 22 per cent of iOS users who had not switched platforms cited iOS security features and 18 per cent cited privacy features. Proposals that push for re-engineering a model – adding a second payment architecture, opening unauthenticated transactions on locked devices to third parties and constraining its abilities to police access – asks consumers who chose integration to bear the risks of openness. Under the CMA’s own framework, a CR must be effective, no more onerous than necessary and must not produce disadvantages disproportionate to its aim. Unfortunately, the measures assigned in the CR fail that test.