By Bronwyn Howell

A defining feature of the digital economy is that information is a non-rival good. Its use by one application does not diminish its potential to generate benefits from another. A second feature is that information exhibits network effects. A single piece of information becomes more valuable when combined with other pieces—as evidenced by the burgeoning use of algorithmic learning.

These features have paved the way for the creation of giant internet platforms such as Amazon, Apple, Google, and Meta. These firms have leveraged information in ways that have unleashed considerable innovation and benefits for consumers. For example, Google gathered information from its search engine, augmented by algorithmic insights, enabling it to enter the market for digital maps with a superior product. Map data has enabled advances in autonomous driving technology, and other applications enabled the firm to enter the health insurance market.

While data-driven innovations have led to economic efficiencies and enhanced consumer experiences, they have also fed concerns about economic dominance, which stifles competition and can lower the incentives to innovate. These concerns underpin calls on both sides of the Atlantic for antitrust intervention to break up big platforms. However, such a blunt structural instrument could create more harm than it averts. Hence, European regulators have proposed two fundamentally distinct policies to keep digital markets competitive and avoid long-run monopolization due to data-driven network effects stemming from algorithmic learning capabilities: mandated “data siloing” (also referred to as data separation) and “data sharing.”

Mandated data siloing prohibits the data-rich incumbent from exploiting consumers’ data gathered in its primary (monopolized) market to improve algorithms and services in a secondary (competitive) market. Theoretically, this would dampen the spillover of data externalities and lower inter-market leverage of data advantage. Data siloing has already been imposed in a number of jurisdictions—for example, when the European Commission approved the merger between Facebook and WhatsApp in 2014—and is also a proposal by the British Competition and Markets Authority. However, data siloing reduces the efficiencies associated with the data-driven network effects between markets.

By contrast, mandated data sharing allows the regulated firm to use its data acquired in the primary market also in the secondary market, but it requires data sharing with rivals so that they compete on more equal footing. Mandated data sharing immediately exploits the non-rivalry of data. Thus, in contrast to data siloing, which limits the extent of data externalities, algorithmic learning, and its associated efficiencies, data sharing permits these efficiency spillovers to the secondary market at the cost of increased competition arising from sharing these efficiencies with rivals. The specific context of click-and-query data from search engines is included under Article 61 of the European Commission’s Digital Markets Act. Yet while mandated data sharing increases the level of competition in the secondary market, it lowers the incumbent’s incentive to innovate in the primary market.

While data siloing and data sharing have regulatory appeal for addressing competition concerns, their impacts on incentives for innovation have not been fully explored. Nevertheless, a recent University of Passau paper seeks to tease out the effects. The paper is notable for the ways it takes into account the size of the data and the strength of externalities available.

The key finding is that data externalities from the primary market not only benefit consumers in the secondary market but also encourage innovation in the primary market. Increased innovation in the primary market expands demand in that market, which then fuels the virtuous cycle of data-driven network effects and the associated data-driven advantage in the secondary market. Data siloing switches off this data externality, thereby lowering innovation incentives of the dominant platform in both the primary and secondary markets. This can ultimately hurt consumers in both markets. By contrast, data sharing preserves the data externality but takes away the data incumbent’s data-driven advantage in the secondary market, which lowers the incumbent’s incentive to collect data in the primary market. Thus, under a data-sharing regime, the incumbent has less incentives to innovate in the primary market, which also lowers innovation incentives in the secondary market.

While more data sharing benefits consumers
in the secondary market, more data siloing hurts consumers in the secondary
market (as data efficiencies are denied). Depending on the strictness of the
regulations, a combination of data siloing and data sharing regulation can
therefore lower consumer surplus, either in the primary market or both markets.
Even if the welfare effect in the secondary market is positive, the unambiguously
negative welfare effect in the primary market can be stronger than the positive
welfare effect in the secondary market.

The message for regulators is clear: Simple
remedies are more complex than they appear at first blush.

The post The Innovation Implications of Mandated Data Siloing and Sharing appeared first on American Enterprise Institute – AEI.