When it comes to information markets, it can seem like we are in the Wild West. There are few rules or regulations in place, and their implementation is far from transparent and settled. Yet the amount of data is growing, the nature of information markets is rapidly changing, and there is considerable potential for opportunity. Only a few years ago, the Federal Trade Commission (2014) could list a few large data brokers; today we have moved to an environment where every large retailer or online service provider – basically anything you sign up for – is a potential trade in the market for data, where personal information can be directly sourced, packaged, and resold.
Information is not only sold directly, but also indirectly in the form of customised goods and services (a distinction first introduced in Admati and Pfleiderer 1990). When you conduct a Google search, for example, advertisers buy a slot on a keyword search results page. Their objective is, of course, to show their link to interested consumers. Thus, by granting access to a targeted consumer population, search engines are de facto bundling advertising slots with information about individual users.
Both pathways of trading information raise privacy concerns, as individuals often do not know – much less control – the ways their data are used. Look at the genetic testing company 23andMe, which partners with the pharmaceutical company GlaxoSmithKline, sharing some of its consumers’ genetic data to develop medical treatments. The structure of information markets – for example, the degree of consumer awareness and information transparency – is bound to affect the availability and security of individual-level information. In turn, privacy concerns will shape the types of data transactions that take place (Bonatti and Cisternas 2018).
Recently, we have seen a few novel attempts to regulate how data is traded. The EU General Data Protection Regulation (GDPR), which went into effect this past May, assigns individuals rights over their data. It addresses how data are managed and ensures that consent management processes are in place. Another example is the recently passed California Consumer Privacy Act, which gives individuals the right to know what data are being collected, why companies collect that data, and how that data are shared. Consumers can tell companies to delete their data or not to sell or share their data.
However, before we get to the point of creating and assessing complex new rules and regulations, we need a much better understanding of how information markets operate, and how they may affect the parties involved.
In a recent survey, we provide a comprehensive model of data trading and brokerage (Bergemann and Bonatti 2018). In the simplest possible scenario, a data broker buys pieces of information from consumers about their taste or willingness to pay, and sells this data to a seller who uses that information to set prices. This seems like a win-win situation for all parties. However, when consumers’ tastes are positively correlated – as in the case of connected nodes in a social network – the marginal value of each consumer’s information (and hence, their equilibrium compensation level) decreases. Consumers lose in this trade and brokers win even if total surplus were to decrease because of data intermediation.
Our work highlights three aspects of information markets, an understanding of which would be beneficial for society, as we continue to engage in policy discussions about regulations.
Value of information
Overall, we know far more about how to sell a given dataset than about how to source data and repackage it as information – for example, in the form of predictions. We have a few specific examples of a data broker setting the price for a consumer’s data. Yet we lack a framework to predict how much money intermediaries, firms, and consumers can really make through an information market. For example, the recent work by Bergemann et al. (2015) indicates how much scope information provides for price discrimination. Beyond bilateral trade, consider the market for credit-score data – even if the value of credit score information for a lender may be transparent, what could be the incentives for businesses to share their databases with Equifax, to contribute to credit scoring in the first place? What compensation do they require?
How does the mode and intensity of competition among buyers of consumer-level information affect its price? What are the implications for consumers? What are the dynamics of competition in information provision, and how does competition among heterogeneous data providers enable firms to better segment their customer populations?
We need to understand the incentives for consumers to reveal data in transactions. How do consumers’ perceptions of the possible uses of their information affect their willingness to engage in data-intensive transactions? Even if it is clear that the consumer owns the data, how will their willingness to sell it depend on the sectors, firms and industries that want to buy it? For example, suppose Amazon wants to access all of your information through a data bank – will you receive better product recommendations? If you appear to be wealthy, will you find only the most expensive brands at the top of your search results? Amazon may have to make it worth your while to reveal your data by providing great recommendations. In the future, it may even have to pay you in cash. In recent work, together with a co-author, we analyse how to design and price information when there is massive heterogeneity among the buyers for data, and the data is high-dimensional (Bergemann et al. 2018).
We have only had limited experience with these rapidly developing markets, and yet huge changes are on the horizon over the next decade. Whatever frameworks we propose today may be less relevant tomorrow. Likewise, any rules we implement today may in some sense influence the development of the marketplace, often in ways that are difficult to anticipate.
These are exciting days, but we would be wise to invest some time and research in better understanding the mechanisms by which information is sold, and the implications of value, competition, and incentives on information markets.
Admati, A and P Pfleiderer (1990), “Direct and indirect sale of information,” Econometrica 58: 901–928.
Bergemann, D and A Bonatti (2018), “Markets for information: An introduction,” Cowles Foundation for Research in Economics Discussion paper.
Bergemann, D, A Bonatti and A Smolin (2018), “The design and price of information,” American Economic Review 108: 1–45.
Bergemann, D, B Brooks and S Morris (2015), “The limits of price discrimination,” American Economic Review 105: 921–957.
Bonatti, A and G Cisternas (2018), “Consumer scores and price discrimination,” MIT, Discussion paper.
Federal Trade Commission (2014), “Data brokers: a call for transparency and accountability”, FTC report.