PI: Prof. Michal Gal.
Research Students: Tamar Indig, Chen Komisar, Lior Shachar, Saar Ben David, Ilana Atron.
Cyberspace, it was hoped, would increase competition. Moreover, it was envisaged that cybernetic competition would significantly increase innovation. The logic is straightforward: transparency of online offers and the increased connectivity between suppliers and consumers makes it more difficult to compete over price, given that if a supplier sets a higher price than its rivals for a similar widget, then consumers can easily switch to buying from them. Therefore, suppliers would have stronger incentives to compete over other features of their offers, mainly the unique and innovative features of their products and services. This, in turn, would increase innovation, thereby also potentially increasing welfare.
These potential benefits of cyberspace are, however, currently threatened. Interestingly, the same traits on which the hopes for a welfare-enhancing cybernetic marketplace were based, have led to the creation of barriers to competition and innovation. The speed of decision-making and its sophistication, coupled with the transparency and connectivity of cyberspace markets, led to a situation in which coordination among competing online suppliers to set a joint profit-maximizing price is easier and more stable than ever before. Indeed, coordination no longer requires competitors to operate in oligopolistic markets; and suppliers can much more quickly and easily detect and punish deviations from the status-quo, thereby reducing incentives for shirking. As our assumptions about which market conditions must exist for firms to coordinate their conduct are altered, the number of red flags that are raised across industries rises. As Ezrachi and Stucke write, this is the end of competition as we know it.
Furthermore, cyberspace make consumers more vulnerable to manipulations by suppliers. Such manipulations can take many forms. One example involves real-time manipulations of emotions. As Facebook recently demonstrated in a controversial experiment on emotional contagion, cybernetic providers may shape the way we feel. Another example involves the manipulation of data. Cyberspace exponentially increased the ability to collect big data. This ability is further reinforced by the advent of the internet-of-things (IOT) and the connectivity of many “things” through cyberspace. The analysis by suppliers of big data regarding consumers’ preferences threatens to strengthen their ability to engage in perfect price discrimination (what some call like to call in the more obscure name “personalized pricing”). Such price discrimination can significantly harm consumer welfare, at least under certain circumstances, by increasing prices charged from consumers to the highest price they are willing to pay for each and every product or service.
These threats to competition and innovation are further increased by the fact that currently some major cyberspace platforms suffer from a high level of concentration. A handful of digital intermediaries with mega platforms control effective points of access to potential users. These include smart devices (iPhone and Kindle), operating systems (iOS and Android), application stores (Apple Store and Google Play) and browser entry points (Google Search and Facebook). The high level of concentration is largely due to network effects, created when the value for each consumer of using the platform rises in parallel with the number of others using the system. These network effects are further increased by the network effects of big data. By converging control of content, access, and online distribution channels, large networks enjoy inherent competitive advantages in access to an immense volume of users’ personal online data. This situation has several implications for welfare. Most importantly, access of other firms to data and to consumers may be affected by the incentives of such intermediaries.
Can the market devise its own solutions to limited competition and innovation in cyberspace? The answer is a partial yes. One potential (albeit partial) solution is the use of algorithms by consumers to counteract at least some of the competition and innovation-reducing conduct of suppliers. Algorithmic consumers (“digital butlers”) are algorithms employed by consumers, which make and execute decisions for the consumer by directly communicating with other systems through the Internet. The algorithm automatically identifies a need, searches for an optimal purchase, and executes the transaction on behalf of the consumer. As elaborated elsewhere, algorithmic consumers offer many benefits to consumers as they can significantly reduce search and transaction costs, and help consumers overcome biases and enable more rational and sophisticated choices.
Most importantly for our purposes, algorithmic consumers can counteract at least some of the negative welfare effects. How can they do so? Algorithmic consumers can create buyer power, if an algorithmic consumer has a sufficiently large number of users, or if it coordinates its conduct with other algorithmic consumers. This, in turn, may allow consumers to counteract suppliers’ buyer power. Indeed, the algorithm can be coded not to buy a certain good if price is above a certain level. The aggregation of buyers can also make transactions less frequent and small, thereby increasing incentives of suppliers to deviate from the status-quo.
Furthermore, algorithmic consumers can be coded to include decisional parameters designed to eliminate or at least reduce some market failures in the long run. Algorithms are sufficiently flexible to include considerations such as long-run effects on market structures that might harm consumers. For example, an algorithm might be able to recognize the coordination, and refrain from doing business with those suppliers until prices are lowered. Or it might always buy some portion of its goods from at least one new source, to strengthen incentives for new suppliers to enter the market. Of course, including such decisional parameters requires more sophisticated modeling and analysis of market conditions and their effect on welfare, but given advances in economics and in data science, they are becoming easier.
Finally, Algorithmic buying groups may reduce the ability of suppliers to learn about, or to use to their advantage, information regarding each user’s preferences by aggregating the choices of different consumers into one virtual buyer (what might be called anonymization-through-aggregation). Indeed, once consumers are aggregated into sufficiently large consumer groups, suppliers will lose the ability to collect information on consumers’ individual preferences with regard to products bought through the group, and to discriminate among them based on each consumer’s elasticity of demand. The loss of this ability, in turn, could increase consumers’ welfare, if suppliers are forced to set a lower price for all.
Algorithmic consumers can therefore improve market dynamics and limit coordination without a need for legal intervention. Rather, their regulating power resides in the reaction of consumers to the change in market conditions created by suppliers, through their algorithms.
This, in turn, will most likely lead to algorithmic wars in cyberspace. Suppliers will not sit quietly while watching some of their benefits from operating in cyberspace being taken away. They will most likely attempt to block access of algorithmic consumers to important inputs (such as data) or to outputs (such as reaching consumers). One strategy that is already observable is for suppliers or cyberspace platform firms to create their own algorithmic consumers, which are not necessarily benign (such as Alexa, Siri, etc.). Indeed, such firms currently dominate the market for algorithmic consumers.
This requires us to think seriously about whether our regulatory tools are fit to the task of ensuring that algorithmic wars in cyberspace increase welfare through creating incentives for competition and innovation. This is the goal of this research. To meet this goal, the research will be conducted in four steps. The first will identify the conditions for competition and innovation in cyberspace, with a special focus being placed on how algorithms operate and on entry barriers into relevant markets. The second will explore in detail the dynamics of algorithmic wars in cyberspace and their effects on competition and innovation. Several scenarios will be explored, which will be sensitive to the special characteristics of the cybernetic marketplace. These two steps will set the stage for the following ones, since without an in-depth understanding of how our markets work, we cannot apprehend the regulatory challenges before us. The third step will analyze existing market and regulatory solutions, to determine their effects on welfare. The final step will involve an exploration of the use of new regulatory tools in order to ensure that algorithmic wars in cyberspace bring about the positive welfare effects they have in store. The methodology includes, inter alia, a thorough literature review, an empirical study of competition law and other regulatory tools employed around the world in order to deal with threats to competition and innovation in cybernetic marketplaces; and the development of a theoretical model for market dynamics in a cyberspace populated by algorithms and characterized by big data and fast connectivity.