The Trophy-Sharing Cartel: How Europe’s Digital Markets Act Turns Google’s IP into a Public Utility

Alden Abbott recently captured Europe’s digital-policy contradiction with a memorable image: Brussels says it wants technology champions, but when a champion emerges, it makes the winner share the trophy. The European Commission’s July 16 binding specification measures (BSMs) against Google turn that metaphor into an administrative program.

One measure requires Google to provide eligible search engines—including AI chatbots with search functions—access to anonymized ranking, query, click, and view data. The other requires “free and effective” interoperability with 11 categories of Android functionality, including device sensors, app context, app and operating-system control, on-device AI models, and background execution. Google must furnish documentation and technical assistance and extend access to covered future functionality when it becomes available to its own services.

These are not conventional rules against fraud or adjudicated anticompetitive conduct. The Commission itself said the specification proceedings “do not take a position on compliance” with the Digital Markets Act (DMA). Yet they produced binding commands reallocating access to assets Google built and continues to maintain.

The measures therefore replace a natural-rights baseline—an owner controls the fruits of its labor absent a legally sufficient justification—with a public-utility baseline: once a platform becomes sufficiently valuable, regulators may redesign it as common infrastructure for competitors. Europe has appointed itself product manager, licensing authority, and rate setter for American intellectual capital.

The Data Commons Fallacy

The Commission treats Google Search’s scale-created advantages as though they were naturally occurring resources that Google happened to fence off. But the organized stream of queries, rankings, clickbacks, dwell signals, fraud filters, metadata, and quality feedback exists because Google invested for decades in crawling, indexing, computing infrastructure, cybersecurity, experimentation, and product design.

Precision matters. Google does not acquire a natural property right in every fact about every user merely because a query passed through its servers. Users retain privacy and other legitimate interests, and many individual facts are not copyrightable or protectable as trade secrets. But it does not follow that a confidential, continuously curated database—and the technical system that generates, cleans, secures, and interprets it—is ownerless. The Commission’s own anonymization, ring-fencing, audit, and security requirements confirm that this is a sensitive, engineered information asset, not an inert public commons.

Randolph May and Seth Cooper locate American intellectual-property law in the Lockean-Madisonian tradition. A person has a right to the fruits of his or her faculties and labor; government is instituted to secure “property of every sort,” not to reassign it to politically favored users. Madison’s “larger and juster meaning” of property expressly encompassed intangible interests. Natural rights supply the grounding principle for IP, even as positive law defines its precise boundaries.

The principle reaches beyond formal patents and copyrights. Trade secrets, confidential know-how, databases, APIs, and integrated software architectures receive different legal protections, but they reflect the same moral baseline: the bread labor has earned belongs presumptively to the laborer. Adam Mossoff likewise warns against converting private IP rights into “regulatory entitlements” whose content is continuously rewritten by administrators. American legal history treated patents as private property secured against government use without license or compensation, not as privileges bestowed by an executive sovereign.

In Ruckelshaus v. Monsanto, the Supreme Court recognized that trade secrets can constitute property and that government-authorized disclosure can implicate the Takings Clause when it defeats legitimate expectations of confidentiality. The U.S. Constitution does not bind Brussels, but the case identifies what the Commission’s vocabulary obscures: compelled access to commercially valuable information is a property rights question, not merely a question of “fairness.”

The Commission answers that it is not transferring Google’s source code or ranking algorithm. True—but incomplete. In a feedback-driven system, valuable know-how is embodied not only in code but in the observed performance of that code across billions of interactions. The Commission itself says recipients may use the mandated data to improve query understanding, ranking, retrieval, and indexing. Access need not reproduce Google Search wholesale to reduce rivals’ costs of experimentation and imitation.

Property includes the right to exclude, choose counterparties, and decide whether, when, and on what terms an asset will be licensed. Leaving Google with a copy while awarding competitors a state-created access right still diminishes ownership.

Dismantling Complementary Assets

The Android mandate is even more revealing because it directs the internal architecture of a complex system.

The Commission requires access to microphone, camera, screen, speaker, location, and sensor inputs; centralized app data; screen automation; operating-system functions; on-device models; computing resources; and background execution. Interoperability must be equally effective, free of charge, supported by documentation and assistance, and extended to covered future improvements. Although the Commission recognizes that several features are sensitive enough to justify certification, it confines restrictions to “exceptional circumstances” and places implementation responsibility on Google.

This is compelled remodularization: the state redraws the boundary between a proprietary core and externally accessible components. It may create more rivalry inside Android, but it can weaken rivalry among integrated systems and suppress experimentation with alternative architectures.

Alchian’s and Allen’s analysis of team production explains why. Complex products are joint outputs. Design, security, hardware, software, testing, distribution, and reputation operate as a coordinated team. The owner of firm-specific assets becomes the residual claimant and therefore receives authority to select participants, monitor performance, and bear gains and losses. Ownership ties control to responsibility.

The Android measure severs that link. Google remains responsible for system integrity, updates, support, and the reputational consequences of failure, while Brussels reallocates significant control rights to third parties. Rivals receive access to complementary assets; Google remains the residual risk-bearer. That is regulatory unbundling without market negotiation.

Richard Langlois’s work on technological systems makes the competitive cost clearer. Competition occurs within modular systems, but also among systems with different degrees of openness, integration, security, and control. That intersystem rivalry gives consumers differentiated choices and lets entrepreneurs test different technical solutions. Antitrust and industrial policy are poorly equipped to choose the optimal boundary because doing so demands immense, dispersed, and rapidly changing knowledge.

This is the Austrian knowledge problem in concrete form. Engineers balance latency, battery life, privacy, permission design, malicious behavior, app compatibility, model performance, and future features that do not yet exist. Those tradeoffs are discovered over time through experimentation and consumer choice. A regulator that mandates “equally effective” access across future versions freezes one conception of fairness into the architecture.

Security magnifies the problem. Pathways allowing an assistant to observe screens, act inside applications, run in the background, invoke sensors, and alter settings necessarily expand trusted relationships and potential failure points. Consent prompts and audits may reduce risk, but cannot abolish it. Nor can Brussels know whether Google’s internal standard and the appropriate standard for an untested third party are genuinely identical. Treating differentiated caution as presumptively discriminatory may encourage firms to delay or simplify advanced functions for everyone.

A Subsidy for Downstream Free-Riders

The search-data pricing rule completes the conversion of property into public utility. Google may generally charge only the incremental cost of preparing, storing, and transmitting the dataset, plus a return on incremental capital capped at its weighted average cost of capital. The measure excludes sunk costs and investments in collecting, processing, and storing the underlying data when they are not attributable specifically to sharing it.

For an information good, that formula is backwards. Digital assets often require enormous fixed investment and uncertain experimentation, while an additional copy may be cheap to deliver. Pricing access near the marginal cost of copying treats creation as economically irrelevant once it succeeds. It compensates the courier but not the creator.

Austrian economics rejects the premise that administrators can derive the “fair” price of an entrepreneurial asset from an accounting formula. Market prices emerge from voluntary exchange and competing judgments about uncertain future uses. They incorporate scarcity, option value, risk, foregone alternatives, and possible failure. A weighted-average-cost-of-capital calculation cannot capture entrepreneurial uncertainty or the value of preserving exclusivity for the next product generation.

Jonathan Barnett’s work on IP licensing is especially important. Strong, enforceable IP often makes specialization possible. An upstream innovator can disclose and license technology to downstream manufacturers and complementors because enforceable rights reduce expropriation risk. Licensing revenues finance further research and development, while voluntary contracts spread technology across an ecosystem. Weakening control can produce the opposite result: firms retreat into integration and secrecy because exchange is no longer safe.

A compulsory “license” priced by the cost of delivery is not an extension of that market. It displaces it. The owner cannot refuse, select partners, protect the option value of exclusivity, or bargain over the asset’s contribution to the recipient’s business. The regulator socializes part of the return while leaving the originator with the cost of generating, maintaining, anonymizing, and defending the asset.

Burtis and Kobayashi show where that logic can lead in software markets. When imitation is substantially cheaper than original creation and IP or contractual protection is too weak, firms rationally choose imitation; the equilibrium supporting investment in original source code can disappear. Static analysis celebrates immediate entry and lower access prices. Dynamic analysis asks the prior question: who will incur the fixed cost of creating the next asset?

The BSM tilts incentives toward an imitator economy. The loss will not appear in a compliance report. It will appear as experiments not funded, datasets not assembled, features not launched in Europe, and architectures designed defensively to minimize what regulators can later compel their creators to share.

Property Without Due Process

The Fifth Amendment is not a cause of action against the European Commission, and the BSMs are reviewable in European courts. The constitutional comparison nevertheless exposes two distinct defects: compelled redistribution on administered terms, and no prior merits determination by an independent adjudicator.

The Commission opened the proceedings, defined the obligations, consulted interested parties, evaluated the technical record, and issued the binding specifications. It did so without finding that Google violated the DMA. By its own description, that was not the proceeding’s purpose. A consultation is not an adversarial trial. Later review is not the same safeguard as requiring the government to prove a violation before a neutral court before redesigning private property.

The American Google search case, litigated before an Article III court, offers a revealing contrast. Whatever one thinks of its outcome—including its own data-sharing remedy—the government first proved unlawful monopolization in a nine-week bench trial. The court then conducted a separate 15-day remedies trial.

In Verizon Communications v. Trinko, the Supreme Court warned of the “uncertain virtue of forced sharing” and of making courts central planners of price, quantity, and terms. In SEC v. Jarkesy, the Court again emphasized constitutional limits on agency adjudication when government seeks punitive civil relief. Neither case controls European law. Together, however, they reflect an American constitutional instinct that coercive government intervention in private rights should be adjudicated, bounded, and institutionally separated from prosecution.

Thomas Sowell warned that political “surrogates” have incentives to deliver concentrated benefits while dispersing costs and obscuring the knowledge they lack. That describes the BSM regime. Selected rivals receive visible access; consumers are promised visible choice. The invisible costs—reduced investment, homogenized architectures, security exposure, and innovations never attempted—are borne diffusely in the future.

A principled defense of IP cannot depend on whether one admires Google. Property rules worthy of the name are neutral. They protect today’s disfavored owner because tomorrow’s innovators (i.e., inventors, authors, entrepreneurs, and startups) need assurance that success will not convert their assets into regulatory entitlements for others.

Europe cannot manufacture champions by redistributing trophies. It can only teach innovators that the safest strategy is not to win—or at least not to win where Brussels can reach the prize.

Image Source: Deposit Photos
Image ID: 6496641
Copyright: stuartmiles