The American Innovation and Choice Online Act
Unlocking Access: The Question of Data Externalities and Potential Costs Associated with Unrestricted Data Access
The American Innovation and Choice Online Act (“AICO”), sponsored by Senator Amy Klobuchar and Senator Chuck Grassley, and co-sponsored by additional bipartisan senators, advanced to the Senate Floor earlier this year.
However, several members of both parties have expressed concerns about the bill and it has yet to be enacted into law. AICO covers the largest online platforms, such as Amazon and Facebook, and makes certain discriminatory conduct by these platforms unlawful—including conduct that materially restricts business users from accessing data generated by the platform. The platform’s use of data to offer competing services also is at issue. However, mandating that platforms cannot restrict business users from accessing data and restricting platforms from using data may lead to less efficient outcomes than those determined by the market.
AICO defines the largest online platforms as those with at least 500,000 monthly consumers or 100,000 business users, market capitalization greater than $600 billion dollars, and that are “a critical trading partner for the sale or provision of any product or service offered on or directly related to the online platform.”
AICO proposes to ban actions by platforms that “preference,” “limit,” or “discriminate” in ways that would “materially harm competition,” conduct that would “materially restrict,” “impede,” or “unreasonably delay” business users from accessing data generated by the platform regardless of whether harm to competition could be shown; and using non-public data that preferences a platform’s own products that compete with a business user’s products regardless of whether harm could be shown. AICO also includes several affirmative defenses. These defenses include showing that the conduct is necessary to comply with other laws, or that the conduct was necessary to protect user safety and privacy or to maintain the core functionality of the platform. For an affirmative defense, the covered platform must also show that the conduct was either “narrowly tailored” or did not result in material harm as shown by the preponderance of the evidence.
AICO does not define “data.” Data likely at issue are data generated or used on the platform, such as consumer identifying information, past purchase history, and other characteristics that would help businesses better design and market products. Data are generated when consumers interact with the platform or one of the businesses on the platform. The data generated through platform interactions may create externalities. Externalities are indirect costs or benefits to third parties who are not directly involved in the economic activity. For example, consumers make shopping decisions that reveal their demand patterns. Online platforms can match these demand patterns to consumer characteristics using machine learning and other business intelligence tools to learn about consumer preferences. This additional learned information is the externality at issue—the platform can now better understand the preferences of new consumers. Because both firms and consumers benefit from a better understanding of shopping patterns, this externality is a positive externality.
Under AICO, these data, and the learning that accrues from the data, must be shared with businesses without restrictions, and the use of these data by the platform would be limited. What this means, for example, is that platforms like Amazon would be unable to use a customer’s past purchasing decisions to offer similar Amazon products to customers. Additionally, platforms like Amazon would not be able to sell customer data to firms or stop firms from accessing their customer data. Instead, Amazon’s customer data would have to be provided to firms free of charge. Additionally, if a platform cannot sell data to businesses or use data to generate its own products, the platform may choose to not collect or generate data. This may result in less efficient service by platforms, because they no longer will generate the learning associated with consumer demand patterns and preferences. Alternatively, if platforms continue to collect and generate data, consumers may be less willing to share the personal data needed for the platform to learn (since the platform, under AICO, must allow unrestricted access to its data).
An alternative to regulatory bills like AICO is to let markets determine how data are shared among consumers and business users. Whether or not markets on their own will ensure consumer data are shared optimally is the focus of several recent economic studies. Rosella Argenziano and Allesandro Bonatti (“Data Linkages and Privacy Regulation,” 2021 working paper) find that consumers and firms benefit when firms share information on past purchases by consumers. Firms benefit because they are able to use that information to offer products and prices in the future that better match the marketplace. Consumers benefit because the new products and prices offered by firms better match their tastes and budgets. Additionally, Dirk Bergemann, Allesandra Bonatti and Alex Smolin (“The Design and Price of Information,” American Economic Review, January 2018) find that different firms want different amounts of data from platforms. Well-informed firms do not need every byte of data, whereas less informed firms desire more information. In equilibrium, a platform will optimally offer a menu of data to firms. Finally, Dirk Bergemann, Allesandro Bonatti, and Tan Gan (“The Economics of Social Data,” RAND Journal of Economics, April 2022) show that platforms will optimally sell information to businesses when it is socially beneficial to consumers and firms to do so, and will not sell consumer’s data when the net benefit to consumers and firms is negative. That is, if consumers value their privacy for some data more than businesses value this data, the online platform will not share these more sensitive consumer information with business users.
Consumer data is a valuable input for the machine learning algorithms that power online commerce. AICO proposes to level the playing field by regulating data access and data use by large online platforms. However, the proposed regulations in AICO also may result in unintended costs, such as a reduction in competition and less efficient online services. Recent economic studies suggest that letting the market determine how online data are shared may result in consumer data being shared optimally in ways that benefit both consumers and businesses.