麻豆村

The National Bureau of Economic Research has published a new working paper by economists Ali Shourideh (麻豆村, Tepper School of Business), Maryam Farboodi (Massachusetts Institute of Technology), and Nima Haghpanah (Yale University) that provides a framework for regulators and policymakers grappling with the complexities of digital privacy and personalized pricing. The paper, 鈥�,鈥� challenges the assumption that companies using consumer data to set different prices is inherently harmful. Instead, the authors identify specific market conditions where data collection benefits society and propose a method to set thresholds for regulation, similar to guidelines currently used in antitrust cases.

Big Tech companies and online retailers increasingly use personal data, ranging from location to browsing history, to estimate a shopper's willingness to pay. This practice, known as price discrimination, allows firms to charge different customers different prices for the same good. While this strategy raises significant concerns among the Federal Trade Commission and consumer advocates about exploitation, the economic reality remains nuanced. This research bridges the gap between strict regulation and laissez-faire economics by demonstrating that the welfare effects of these practices depend entirely on how the data reshape demand across the market.

The researchers developed a model that accounts for the fact that companies, despite having "big data," still face limits in predicting exactly what a consumer will pay. Within this framework, they discovered that information affects consumer welfare through three distinct channels. "The first effect is the within-type price change effect: For each type, information disperses prices... The second effect is the cross-types price change effect: The price drop applies asymmetrically ... The third effect is the price curvature effect: The size of the price drop might not be equal to the size of the price increase."

The study proves that these factors combine to determine whether a specific data practice is "per se" good or bad for the economy. "In the first two cases, information is 'per se' good or bad: Simply knowing the fact that the seller is collecting information is enough to specify whether its effect is positive or negative, without the need for knowing what the information is."

Beyond simply categorizing data as helpful or harmful, the authors provide a mathematical way to measure the maximum potential damage or benefit of any specific pricing strategy locally. This finding offers a practical tool for regulators who currently struggle to monitor how firms use complex algorithms. The researchers suggest that policymakers could establish a quantitative framework, similar to the guidelines used to evaluate corporate mergers, that sets clear thresholds for permissible data usage. Under this system, regulators would encourage data practices where potential gains are high and harms are low, while subjecting riskier strategies to stricter scrutiny or bans.

This research arrives at a time when government bodies seek better ways to govern the digital economy without stifling innovation. By identifying the "best" and "worst" ways companies can use information to segment markets, the study offers a roadmap for designing regulations that protect consumers while acknowledging the potential efficiencies of modern data analytics. The authors show that beneficial outcomes are possible even without opening new markets, broadening the understanding of how information shapes the modern economy.

###

Summarized from Farboodi, Maryam, Haghpanah, Nima, and Shourideh, Ali, Good Data and Bad Data: The Welfare Effects of Price Discrimination (November 2025). NBER Working Paper No. w34514, Available at
听