Increasingly, algorithms are supplanting human decision-makers in pricing goods and services. To analyze the possible consequences, we study experimentally the behavior of algorithms powered by Artificial Intelligence ( Q-learning) in a workhorse oligopoly model of repeated price competition. We find that the algorithms consistently learn to charge supracompetitive prices, without communicating with one another. The high prices are sustained by collusive strategies with a finite phase of punishment followed by a gradual return to cooperation. This finding is robust to asymmetries in cost or demand, changes in the number of players, and various forms of uncertainty. (JEL D21, D43, D83, L12, L13) Software programs are increasingly being adopted by firms to price their goods and services, and this tendency is likely to continue. 1 In this paper, we ask whether pricing algorithms may "autonomously" learn to collude. The possibility arises because of the recent evolution of the software, from rule-based to reinforcement learning programs. The new programs, powered by Artificial Intelligence (AI), are indeed much more autonomous than their precursors. They can develop their pricing strategies from scratch, engaging in active experimentation and adapting to changing environments. In this learning process, they require little or no external guidance.In the light of these developments, concerns have been voiced, by scholars and policymakers alike, that AI pricing algorithms may raise their prices above the competitive level in a coordinated fashion, even if they have not been specifically

This paper studies the exchange of information between two principals who contract sequentially with the same agent, as in the case of a buyer who purchases from multiple sellers. We show that when (a) the upstream principal is not personally interested in the downstream level of trade, (b) the agent's valuations are positively correlated, and (c) preferences in the downstream relationship are separable, then it is optimal for the upstream principal to offer the agent full privacy. On the contrary, when any of these conditions is violated, there exist preferences for which disclosure is strictly optimal, even if the downstream principal does not pay for the information. We also examine the effects of disclosure on welfare and show that it does not necessarily reduce the agent's surplus in the two relationships and in some cases may even yield a Pareto improvement.

No abstract

Increasingly, algorithms are supplanting human decision-makers in pricing goods and services. To analyze the possible consequences, we study experimentally the behavior of algorithms powered by Artificial Intelligence (Q-learning) in a workhorse oligopoly model of repeated price competition. We find that the algorithms consistently learn to charge supracompetitive prices, without communicating with one another. The high prices are sustained by collusive strategies with a finite phase of punishment followed by a gradual return to cooperation. This finding is robust to asymmetries in cost or demand, changes in the number of players, and various forms of uncertainty. (JEL D21, D43, D83, L12, L13)

We study the effects of reminders on people’s behavior in investment activities characterized by up-front costs and delayed benefits, such as getting an education and maintaining a healthy lifestyle. We conduct a field experiment and show that simple weekly reminders induce users of a gym to substantially increase their gym attendance over an extensive period. Users’ response to reminders is immediate (within hours) and recurrent for any subsequent reminder, which can be explained by limited attention. We find some evidence of habit formation, leading to more frequent physical activity also after treatment, although with an effect smaller than during treatment. Simple reminders are thus a cost-effective policy tool. Data are available at http://dx.doi.org/10.1287/mnsc.2016.2499 . This paper was accepted by Uri Gneezy, behavioral economics.

We analyze firms that compete by means of exclusive contracts and market-share discounts (conditional on the seller's share of customers' total purchases). With incomplete information about demand, firms have a unilateral incentive to use these contractual arrangements to better extract buyers' informational rents. However, exclusive contracts intensify competition, thus reducing prices and profits and (in all Pareto undominated equilibria) increasing welfare. Market-share discounts, by contrast, produce a double marginalization effect that leads to higher prices and harms buyers. We discuss the implications of these results for competition policy. (JEL D43, D83, D86, K21, L14, L42)

We examine the intricacies associated with the design of revenue-maximizing mechanisms for a monopolist who expects her buyers to resell. We consider two cases: resale to a third party who does not participate in the primary market and interbidder resale, where the winner resells to the losers. To influence the resale outcome, the monopolist must design an allocation rule and a disclosure policy that optimally fashion the beliefs of the participants in the secondary market. Our results show that the revenue-maximizing mechanism may require a stochastic selling procedure and a disclosure policy richer than the simple announcement of the decision to sell to a particular buyer.2 Haile (2003) and Schwarz and Sonin (2001) consider models where bidders' valuations change over time. 3 Although not considered in this article, participation only in secondary markets may also be strategic, as indicated in McMillan (1994) and Jehiel and Moldovanu (1996). 4 See also Gupta and Lebrun (1999) for an analysis of first-price asymmetric sealed-bid auctions followed by resale, where trade in the secondary market is motivated by the inefficiency of the allocation in the primary market. 5 Revenue-maximizing mechanisms without resale have been examined, among others, by Maskin and Riley (1984) and Myerson (1981).

Pricing decisions are increasingly in the "hands" of artificial algorithms. Scholars and competition authorities have voiced concerns that those algorithms are capable of sustaining collusive outcomes more effectively than human decision makers. If this is so, then our traditional policy tools for fighting collusion may have to be reconsidered. We discuss these issues by critically surveying the relevant law, economics and computer science literatures. 5 For example, the New Yorker asked what happens "When bots collude" (Algorithmic pricingril 25 th , 2015), and the Financial Times talked of "Digital cartels" (January 8 th , 2017).