Social Sciences Brown Bag Seminar
Abstract: This paper is closely related to the recent empirical work that studies the interaction of market power and selection in insurance and credit markets. With the rapid development of big data, machine learning, and AI, competing firms may have access to different amounts of data, and use different pricing algorithms. This has raised important questions for competition policy. Our paper develops a method to empirically analyze competitive equilibrium in selection markets when firms offer differentiated products while having different cost structures and information precision. We apply the method to study the Italian auto insurance industry using a panel dataset covering all liability contracts and claims for over a decade. We find that there were substantial differences in the precision of risk rating across firms, but firms that suffer lower information precision tend to have lower costs. Forcing all firms to adopt the worst information technology counterfactually (e.g., privacy regulation) leads to a lower equilibrium market price; consumer surplus increases, mainly for high risk drivers.
Joint with Marco Cosconati, Yizhou Jin, and Fan Wu