Back

Poster Sessions

Monday, July 05 11:45-13:00 EST

Alexander Rodivilov

School of Business, Stevens Institute of Technology

Contract Designers and Incomplete Contracts    

Aram Grigoryan

Duke University

Priority-based Assignment with Reserves and Quotas

(joint work with Atila Abdulkadiroğlu)

Brian C. Albrecht

Kennesaw State University

Market Microstructure and Informational Efficiency

(joint work with Rafael R. Guthmann)

Cheng Guo

University of Toronto

Copositive Duality for Discrete Markets and Games   [Paper]

(joint work with Merve Bodur and Joshua A. Taylor)

Abstract

Optimization models for economic problems with binary decisions are nonconvex and thus lack strong duality, which limits the usefulness of tools such as shadow prices and the KKT conditions. For example, in convex markets, shadow (dual) prices are associated with market equilibrium, and for convex games the existence and uniqueness of Nash equilibrium can be proved via fixed-point theorem and KKT conditions. Those results are lacking in their nonconvex counterparts. We use copositive programming to formulate discrete problems in applications including nonconvex energy markets and nonconvex games, to leverage its convexity and strong duality features. We obtain several novel theoretical and numerical results for those applications, including a new revenue-adequate pricing scheme for energy markets, and existence, uniqueness, and KKT conditions for the pure-strategy Nash equilibrium in discrete games. We also propose a novel and easy-to-implement cutting-plane algorithm for mixed-integer copositive programs, and employ it in our applications.

Clemens Possnig

University of British Columbia

Algorithmic Collusion and Best Response Dynamics for Repeated Games

Abstract

Reinforcement learning agents in their application as pricing algorithms have attracted a lot of attention from economists in recent years. Both empirical evidence and numerical simulations suggest that a market composed of reinforcement algorithms may be vulnerable to collusive pricing. However, the literature still lacks an understanding of how the possibility of algorithmic collusion depends on details of the competitive environment, represented by primitives of the repeated game model used, and features of the learning algorithm. This project will attempt to shed light on this issue. I study the limiting behavior of independent Q-learning agents playing a repeated Cournot game and obtain conditions on the model primitives and algorithm features for collusion to arise. To the best of my knowledge, this is the first study that gives such analytical insights when the learning algorithms considered are independent (i.e. do not communicate) and can condition their policies on the history of the game. The analysis focuses on a special class of Q-learning algorithms known as actor-critic, and extends stochastic approximation techniques to a repeated game with history dependent policies. In so doing, I analyze a novel form of policy updating rule that can be seen as best response dynamics adapted to Markov strategies that is of interest in its own right.

Daniel Schoepflin

Drexel University

Bayesian Clock Auctions

(joint work with Michal Feldman, Vasilis Gkatzelis and Nick Gravin)

Denizalp Goktas

Brown University

Tâtonnement beyond Constant Elasticity of Substitution: A Generalization of Eisenberg-Gale’s Program

(joint work with Enrique Areyan Viqueira and Amy Greenwald)

Eduardo Garcia

University of Rochester

Learning in Games Where Agents Sample

Felix Feng

University of Washington

Looking the Other Way: The Screening Role of (Weak) Internal Monitoring

(joint work with Wenyu Wang and Yufeng Wu)

Francisco Castro

UCLA Anderson School of Management

Mechanism Design under Approximate Incentive Compatibility

(joint work with Santiago Balseiro and Omar Besbes)

Gabriel P. Andrade

University of Colorado Boulder

Learning in Matrix Games can be Arbitrarily Complex

(joint work with Rafael Frongillo and Georgios Piliouras)

Gregory Raiffa

University of California San Diego

Strategic Settings with Social and Psychological Elements

(joint work with Joel Watson)

Jiadong Gu

University of North Carolina at Chapel Hill

Competing for Consumer Attention

Abstract

This paper studies the buyers’ optimal learning and its interaction with sellers’ competition, especially, pricing and product differentiation. Buyers acquire costly information about the match before making trading decisions. Sellers choose prices and product differentiation to attract buyers. In the equilibrium with product differentiation, buyers’ optimal learning induces a Logit-formula demand, and the information cost balances the sellers’ price competition and choices of product differentiation. Instead, in the equilibrium without differentiation, buyers won’t acquire information and sellers involve in the Bertrand competition. In addition, the information cost to buyers could improve the market efficiency as it increases the supply-side market power and decreases the sellers’ incentive to over-differentiate themselves to soften the price competition.

Komal Malik

Indian Statistical Institute, Delhi

Optimal robust mechanism in bilateral trading

Lucía Pinar García

University of Valencia

Internet duopoly two-sided market: A zero rating analysis

(joint work with Iván Arribas and Amparo Urbano)

Abstract

The aim of this paper is to study the effect of zero-rating contracts in an Internet duopoly two-sided market. We propose a two-sided market game, with two Internet Service Providers (ISPs) offering Internet services to a representative consumer (EU) to access to two asymmetric content platforms (CPs). Our model is a three period game and we study the effects of the introduction of zero-rating services with inter-platform competition. We find that the higher capacity cost ISP is never chosen by the consumer. The social welfare quantities are higher than those of the Pareto optimal subgame perfect equilibrium and the CPs are always better off with zero rating than without it.

Mark Whitmeyer

University of Bonn

Search and Competition with Flexible Investigations

(joint work with Vasudha Jain)

OSub Kwon

Ohio State University

Bayesian Persuasion in the Lab

Pedro Vaissman Guinsburg

Universidade de São Paulo

Information Design and Sensitivity to Market Fundamentals

Qiongyuan Cao

The University of Edinburgh

Bilateral Bargaining in Networks

(joint work with Joosung Lee)

Shaul Rosner

The Interdisciplinary Center Herzliya (IDC Herzliya)

Race Congestion Games   [Paper]

(joint work with Tami Tamir)

Abstract

We consider a new variant of congestion games, arising in environments with strong competition. In our setting, denoted race congestion games, the players are partitioned into competition sets, and the goal of every player is to perform well relative to its competitors.

We analyze the race variant of various congestion games. We show that race games are significantly different from classical games -- in which players aim to minimize their cost, independent of the cost of other players. Specifically, competition may lead to poor outcomes: even simple race congestion games may not have a pure Nash equilibrium, best-response dynamics may not converge to a NE even if one exists, and the equilibrium inefficiency is high. Moreover, deciding if a race game has a pure NE is NP-hard.

We identify several non-trivial classes of games for which a NE exists and can be computed efficiently, and present tight bounds on the equilibrium inefficiencies. Striving for stability, we suggest and study Nashification of race games by slight modifications of the instance.

Our analysis considers race scheduling games, which is the race variant of weighted singleton congestion games, as well as race cost-sharing games. Our initial results on race scheduling games were presented in SAGT 2020, and a full version can be found at https://cs.idc.ac.il/~tami/Papers/RSG-full.pdf.

Shuo Xu

The Ohio State University

Information Obfuscation

Tao Zhang

New York University

Informational Design of Multi-Agent System

(joint work with Quanyan Zhu)

Weixuan Zhou

Hong Kong University of Science and Technology

Indirect Sales with Search Cost

(joint work with Feiting Xu)

Ying Gao

MIT

Inference with Selectively Disclosed Data

Wednesday, July 07 11:45-13:00 EST

Andrew Ferdowsian

Princeton University

Build to Order: Endogenous Supply in Centralized Mechanisms

(joint work with Kwok Hao Lee and Luther Yap)

Berk Idem

Penn State

Coexistence of Centralized and Decentralized Markets

Chen Lyu

University of Wisconsin -- Madison

Information Design for Selling Search Goods and the Effect of Competition

Abstract

I study optimal information provision by a search goods seller. While the seller controls a consumer's pre-search information, which decides whether she will search for the product with a cost, he cannot control the consumer's post-search information because the consumer inevitably learns her match upon inspecting the product. Importantly, I allow the consumer to have private information on her best alternative to the seller's product. To accommodate this, a relaxed problem approach is developed, which fully characterizes the seller's optimal design under certain regularity conditions. Based on this characterization, I consider how consumer privacy, revealing of post-search information, reduction in search cost and competition by a large number of sellers would affect the equilibrium information provision. My approach also helps to highlight the key similarity and dissimilarity between information designs for search goods and experience goods.

Chin-Chia Hsu

MIT, IDSS

Verification, Disclosure and Network Formation in News Subscription Services

(joint work with Amir Ajorlou, Muhamet Yildiz and Ali Jadbabaie)

Deepanshu Vasal

Northwestern University

Network design for social welfare   [Paper]

(joint work with Abhishek Shende and Sriram Vishwanath)

Abstract

In this paper, we consider the problem of network design on network games. We study the conditions on the adjacency matrix of the underlying network to design a game such that the Nash equilibrium coincides with the social optimum. We provide the examples for linear quadratic games that satisfy this condition. Furthermore, we identify conditions on properties of the adjacency matrix that provide a unique solution using variational inequality formulation, and verify the robustness and continuity of the social cost under perturbations of the network. Finally we comment on individual rationality and extension of our results to large random networked games.

Duoxi Li

Boston University

Relational Contracts in Frictional Markets with Rematching

Frank Yang

Stanford University

Static Pricing   [Paper]

(joint work with Martino Banchio)

Abstract

A monopolist sells items repeatedly over time to a consumer with persistent private information. The seller has limited commitment: she cannot commit to a long-term contract but always has the option to commit to posted prices for unsold items. A pricing strategy is static if it is history-independent. We show that the seller adopts static pricing in equilibrium. The ratchet effect eliminates potential gains from dynamic pricing for any degree of persistence of the private information.

Georgy Lukyanov

École Polytechnique

Ideological Polarization in the Presence of Fake News

(joint work with Samuel Safaryan)

Ji Hee Yoon

University College London, Economics

Innovation in Market Clearing Technology: Design of Multi-Asset Trading Mechanisms

(joint work with Chen Lyu and Marzena Rostek)

Kartik Ahuja

Montreal Institute for Learning Algorithms (MILA)

Linear Regression Games: Convergence Guarantees to Approximate Out-Of-Distribution Solutions

(joint work with Karthikeyan Shanmugam and Amit Dhurandhar)

Abstract

Recently, invariant risk minimization (IRM) [Arjovsky et al., 2019] was proposed as a promising solution to address out-of-distribution (OOD) generalization. In [Ahuja et al., 2020], it was shown that solving for the Nash equilibria of a new class of “ensemble-games” is equivalent to solving IRM. In this work, we extend the framework in [Ahuja et al., 2020] for linear regressions by projecting the ensemble-game on an L infinity ball. We show that such projections help achieve non-trivial OOD guarantees despite not achieving perfect invariance. For linear models with confounders, we prove that Nash equilibria of these games are closer to the ideal OOD solutions than the standard empirical risk minimization (ERM) and we also provide learning algorithms that provably converge to these Nash Equilibria. Empirical comparisons of the proposed approach with the state-of-the-art show consistent gains in achieving OOD solutions in several settings involving anti-causal variables and confounders.

Konstantin Zabarnyi

Technion - Israel Institute of Technology

Regret-Minimizing Bayesian Persuasion

(joint work with Yakov Babichenko, Inbal Talgam-Cohen and Haifeng Xu)

Marcos Ross Fernandes

Getulio Vargas Foundation

Sudan Conflict: a network approach (Networks, Social Conflict)

(joint work with Samuele Centorrino, Alejandro Melo Ponce and Camilo Rubbini)

Nicolas Pastrian

University of Pittsburgh

Surplus Extraction with Behavioral Types     [Paper]

Abstract

We reexamine the surplus extraction problem in a mechanism design setting with behavioral types. We focus on an extreme class of behavioral types which always perfectly reveal their private information. However, incentive compatibility constraints remain necessary for strategic types with respect to the contract of all types. We characterize the sufficient conditions that guarantee full extraction in the reduced form environment of McAfee and Reny (1992). The standard convex independence condition identified in Crémer and McLean (1988) remains necessary only among the beliefs of strategic types, while a weaker condition is required for the beliefs of behavioral types.

Parth Parihar

Princeton University

Limited Foresight and Gridlock in Bargaining

Pengfei Zhang

Cornell University

The Human Side of Cyber Property Rights: Theory and Evidence from Github

Rachitesh Kumar

Columbia University

Contextual First-Price Auctions with Budgets

(joint work with Santiago Balseiro and Christian Kroer)

Shivika Narang

Indian Institute of Science

On Achieving Fairness and Stability in Many-to-One Matchings

(joint work with Arpita Biswas and Y Narahari)

Silvia Martinez-Gorricho

Universidad Catolica de la Santisima Concepcion

Incentives, Ability and Disutility of Effort   [Paper]

(joint work with Miguel Sánchez-Villalba)

Sulagna Dasgupta

University of Chicago

Information Design in One-sided Matching Markets

Abstract

In the one-sided matching problem, objects are allocated to agents based on agent preferences. However, agents may not always know, a priori, their cardinal or even ordinal preferences over the objects, because they do not have enough information. In this context, I try to answer the question: How should a benevolent planner optimally reveal information to the agents to maximize welfare, in an environment where agents have no private information to start with? As a benchmark, I first show that when using any of the standard strategyproof ordinal mechanisms, such as Deferred Acceptance, Serial Dictatorship, Random Priority or Top Trading Cycle, letting each agent know his true ordinal ranking over the objects is almost never a social welfare-maximizing information policy. By way of a partial solution, I then propose a simple signal I call the Object Recommendation (OR) Signal. Under i.i.d. agent priors satisfying a mild regularity condition, I show that, when agents' a priori relative preferences over the objects are "not too strong", the OR Signal, used together with any of the aforementioned standard mechanisms, not only maximizes welfare, but achieves first-best, i.e. the unconstrained maximum total ex-ante welfare.

Toygar T. Kerman

Maastricht University

Belief Inducibility and Informativeness

(joint work with P. Jean-Jacques Herings and Dominik Karos)

Abstract

We consider a group of receivers who share a common prior on a finite state space and who observe private correlated signals that are contingent on the true state of the world. We show that, while necessary, Bayes plausibility is not sufficient for a distribution over posterior belief vectors to be inducible, and we provide a characterization of inducible distributions. We classify communication strategies as minimal, direct, and language independent, and we show that any inducible distribution can be induced by a language independent communication strategy (LICS). We investigate the role of the different classes of communication strategies for the amount of higher order information that is revealed to receivers. We show that the least informative communication strategy which induces a fixed distribution over posterior belief vectors lies in the relative interior of the set of all language independent communication strategies which induce that distribution.

Yuan Gao

Columbia University

Infinite-Dimensional Fisher Markets and Tractable Fair Division   [Paper]

(joint work with Christian Kroer)

Abstract

We extend the classical Eisenberg-Gale framework for linear Fisher market equilibria to the case of measurable item spaces, allowing an infinite number or a continuum of items. Given such a market, we show that an infinite-dimensional Eisenberg-Gale convex program (EG) and its dual have optimal solutions that correspond to its (infinite-dimensional) market equilibria. For an item space with a finite, atomless measure, we show that there always exists a pure market equilibrium (i.e., a fair division if all buyers have the same budget), in which buyers receive a.e.-disjoint measurable subsets as their equilibrium allocations. An infinite-dimensional market equilibria also possesses desirable fairness and efficiency properties, just as in the classical case of finitely many items. We show that, when buyers have piecewise linear valuations on a closed interval (the item space), the infinite-dimensional EG can be reformulated into a tractable convex conic program and hence solved using off-the-shelf optimization software. This gives a theoretically and practically efficient algorithm for computing a pure market equilibrium in this case. For more general buyer valuations, we propose solving the dual of EG via first-order stochastic optimization to compute equilibrium utility prices and establish convergence guarantees. Finally, we extend the above results to the case of quasilinear buyer utilities.

Back