« Back to Results

Artificial Intelligence, Big Data, and Competition with Algorithms

Paper Session

Saturday, Jan. 4, 2020 2:30 PM - 4:30 PM (PDT)

Manchester Grand Hyatt, Harbor C
Hosted By: Korea-America Economic Association & American Economic Association
  • Chair: Jay Pil Choi, Michigan State University

Counterfactual Inference for Consumer Choice Across Many Product Categories

Susan Athey
,
Stanford University
Robert Donnelly
,
Facebook
Francisco Ruiz
,
Deep Mind
David Blei
,
Columbia University

Abstract

This paper proposes a method for estimating consumer preferences among discrete choices, where the consumer chooses at most one product in a category, but selects from multiple categories in parallel. The consumer's utility is additive in the different categories. Her preferences about product attributes as well as her price sensitivity vary across products and are in general correlated across products. We build on techniques from the machine learning literature on probabilistic models of matrix factorization, extending the methods to account for time-varying product attributes and products going out of stock. We evaluate the performance of the model using held-out data from weeks with price changes or out of stock products. We show that our model improves over traditional modeling approaches that consider each category in isolation. One source of the improvement is the ability of the model to accurately estimate heterogeneity in preferences (by pooling information across categories); another source of improvement is its ability to estimate the preferences of consumers who have rarely or never made a purchase in a given category in the training data. Using held-out data, we show that our model can accurately distinguish which consumers are most price sensitive to a given product. We consider counterfactuals such as personally targeted price discounts, showing that using a richer model such as the one we propose substantially increases the benefits of personalization in discounts. We conclude that richer data as well as richer, machine-learning based models both improve the ability of firms to profit from targeted pricing policies.

Statistical Discrimination in Ratings-Guided Markets

Yeon-Koo Che
,
Columbia University
Kyungmin Kim
,
Emory University
Weijie Zhong
,
Stanford University

Abstract

We study statistical discrimination of individuals based on payoff-irrelevant social identities in markets where ratings/recommendations facilitate social learning among users. Despite the potential promise and guarantee for the ratings/recommendation algorithms to be fair and free of human bias and prejudice, we identify possible vulnerability of the ratings-based social learning to discriminatory inferences on social groups. In our model, users’ equilibrium attention decision may lead data to be sampled differentially across different groups so that differential inferences on individuals may emerge based on their group identities. We explore policy implications in terms of regulating trading relationships as well as algorithm design [to be added].

Competition Law and Pricing Algorithms

Joseph Harrington
,
University of Pennsylvania

Abstract

TBA

Data and Competition

Alexandre de Cornière
,
Toulouse School of Economics
Greg Taylor
,
University of Oxford

Abstract

The question of data has been at the center of recent debates around competition policy in the digital era. Concerns in this area are wide-ranging, and encompass privacy, collusion, barriers to entry, exploitative practices, and data-driven mergers.
Data can serve several purposes: for instance it can be used to improve algorithms, to target advertising, or to offer personalized discounts to consumers. While this heterogeneity of uses for data has sparked a large literature in economics, the multiplicity of models makes it difficult to draw general conclusions about the competitive effects of data.

In this paper we introduce data into a competition-in-utility framework. The three key features of data are that (i) it allows to generate more revenue for a given level of utility, (ii) it is a byproduct of firms' economic activity, and (iii) it is a club good (non-rival and excludable).

We provide a sufficient condition for data to be pro-competitive, and apply it to several environments illustrating the variety of uses for data. We then use the framework to study market dynamics, data-driven mergers and privacy policies.
Discussant(s)
Sokbae Lee
,
Columbia University
Joshua Gans
,
University of Toronto
Ilwoo Hwang
,
University of Miami
Hyojin Song
,
Microsoft Research
JEL Classifications
  • C8 - Data Collection and Data Estimation Methodology; Computer Programs
  • D2 - Production and Organizations