Personalized pricing strategies of competitive firms when consumers can manipulate data to beat AI algorithms
摘要
While firms collect consumer data to train algorithms for price discrimination, there exist sophisticated consumers manipulating their data to beat AI algorithms for better deals. This paper develops a game-theoretical model to capture consumer data manipulation behavior in a duopoly market. Using the model framework, we analyze how consumer data manipulation behavior affects the equilibrium prices, profits, consumer surplus, and social welfare. We find that, when two firms adopt different pricing strategies, consumer data manipulation behavior benefits the firm adopting the uniform pricing strategy due to the competition mitigating effect, but it hurts the rival because of the consumer utility decreasing effect. Besides, it backfires on consumers with higher prices and extra manipulation costs. And it can enhance or reduce social welfare, depending on whether the consumer utility decreasing effect or production cost reducing effect dominates. We also examine how the competitive firms choose between the personalized pricing strategy and the uniform pricing strategy in equilibrium. Our results suggest that, when the proportion of sophisticated consumers is small or the manipulation cost efficiency is large, two firms both adopt the personalized pricing strategy. Otherwise, two firms both adopt one of the two pricing strategies, where they earn higher profits under the uniform pricing strategy. Our findings confirm the validity of personalized pricing strategies even when consumers engage in data manipulation, while also accounting for the continued prevalence of uniform pricing in the era of big data.