Management sciences, decision sciences and quantitative methods; Product and operations management; Technology, R&D management
Data, machine learning, data product, pricing, incentives, contracting
This paper explores how firms that lack expertise in machine learning (ML) can leverage the so-called AI Flywheel effect. This effect designates a virtuous cycle by which, as an ML product is adopted and new user data are fed back to the algorithm, the product improves, enabling further adoptions. However, managing this feedback loop is difficult, especially when the algorithm is contracted out. Indeed, the additional data that the AI Flywheel effect generates may change the provider's incentives to improve the algorithm over time. We formalize this problem in a simple two-period moral hazard framework that captures the main dynamics among ML, data acquisition, pricing, and contracting. We find that the firm's decisions crucially depend on how the amount of data on which the machine is trained interacts with the provider's effort. If this effort has a more (less) significant impact on accuracy for larger volumes of data, the firm underprices (overprices) the product. Interestingly, these distortions sometimes improve social welfare, which accounts for the customer surplus and profits of both the firm and provider. Further, the interaction between incentive issues and the positive externalities of the AI Flywheel effect has important implications for the firm's data collection strategy. In particular, the firm can boost its profit by increasing the product's capacity to acquire usage data only up to a certain level. If the product collects too much data per user, the firm's profit may actually decrease, i.e., more data is not necessarily better.
Copyright © 2022, INFORMS
Health and environment; Information technology and systems
Public health, epidemic control, information design, strategic behavior
This paper explores how governments may efficiently inform the public about an epidemic to induce compliance with their confinement measures. Using an information design framework, we find the government has an incentive to either downplay or exaggerate the severity of the epidemic if it heavily prioritizes the economy over population health or vice versa. Importantly, we find that the level of economic inequality in the population has an effect on these distortions. The more unequal the disease's economic impact on the population is, the less the government exaggerates and the more it downplays the severity of the epidemic. When the government weighs the economy and population health sufficiently equally, however, the government should always be fully transparent about the severity of the epidemic.
© 2021, INFORMS
Management sciences, decision sciences and quantitative methods
Intermediary problems, mechanism design, internet advertising, extensive form games, second-price auction, multi-stage intermediation
We consider a setting where online advertisers seek to acquire impressions from an advertising exchange through a multi-tier network of intermediaries, and study the mechanisms offered by the ad exchange and intermediaries when the advertisers’ values are private. As opposed to traditional manufacturer/retailer settings, intermediaries in display advertising auction off contingent goods which they only purchase when downstream buyers have signaled interest. Thus motivated, we determine how intermediaries should bid on behalf of their customers in the mechanism of an upstream intermediary and study how the structure of the intermediation network affects the profits of its participants. We provide a game theoretic model to study the mechanisms offered by the ad exchange and intermediaries within a practically relevant class of mechanisms. We characterize a subgame perfect equilibrium of the game among intermediaries and the seller, and show that the equilibrium mechanisms have a simple and appealing structure: intermediaries bid the virtual value associated with the maximum downstream report in the upstream intermediary’s mechanism, whenever this quantity is positive. We show that economic incentives are not necessarily aligned along the network and the position in the intermediation network has a significant impact on the profits of the intermediaries. Besides, we analyze the impact of different network structures on the seller’s revenue and investigate whether intermediaries have an incentive to merge horizontally or vertically.
Copyright © 2020, INFORMS
Management sciences, decision sciences and quantitative methods; Product and operations management
Dynamic mechanism design, social efficiency, multi-agent games, resource allocation without money
We consider a principal repeatedly allocating a single resource in each period to one of multiple agents, whose values are private, without relying on monetary payments over an infinite horizon with discounting. We design a dynamic mechanism without monetary transfers, which induces agents to report their values truthfully in each period via promises/threats of future favorable/unfavorable allocations. We show that our mechanism asymptotically achieves the first-best efficient allocation (the welfare-maximizing allocation as if values are public) as agents become more patient and provide sharp characterizations of convergence rates to first best as a function of the discount factor. In particular, in the case of two agents we prove that the convergence rate of our mechanism is optimal, i.e., no other mechanism can converge faster to first best.
Copyright © 2019, INFORMS
Airline scheduling, aircraft fleeting and routing, cruise time controllability, second order cone programming