RESEARCH
I am an economist trained in empirical industrial organization and information economics. I study how data and AI technologies generate useful information in imperfect markets and the mechanisms through which this information creates and distributes economic value.
I focus on applications in insurance and digital platforms. I specialize in structural modeling, field experiments, and independent collaboration with firms.
INSURANCE
Firms in many markets directly elicit large amounts of data from consumers. The data is used to mitigate information problems, gain a competitive advantage, and extract rent from consumers. We develop an equilibrium framework to trade-off these countervailing
forces, and use it to study the creation of detailed driving data in U.S. auto insurance through a voluntary monitoring program. Surprisingly, the incentive effect from the data collection process, not the data itself, is the main source of social value creation. Despite large switching inertia, the product market is still competitive enough so that forcing the firm to make the data public hurts short-term consumer welfare by discouraging data creation.
In macro/finance, our model is typically referred to as "rational inattention:" in a market with asymmetric information, firms endogenously acquire information subject to a cost (switching inertia and consumers' unwillingness to be monitored).
(corresp. author | w/ Shosh Vasserman)
In many economic settings -- e.g. consumer lending, criminal justice, labor contracts -- the intuition of moral hazard has given rise to a doctrine of deterring bad behaviors by increasing punishment. In auto insurance, firms combat risky driving almost exclusively by raising accident punishment. However, not only does this undermine risk-sharing, this paper shows that it fails at deterring risky driving due to drivers' inattention to risk. We do so with novel sensor data as well as observational and experimental methods. Our results illustrate the important interaction between bounded rationality and market failure, and how behavioral data allow insurers to expand risk management strategies from selective risk-sharing to risk mitigation.
This paper studies competition under information heterogeneity in selection markets and examines the impact of public information regulations aimed at reducing information asymmetries between competing firms. We develop a novel model and introduce new empirical strategies to analyze imperfect competition in markets where firms have heterogeneous information about consumers, vary in cost structures, and offer differentiated products. Using data from the Italian auto insurance market, we find substantial differences in the precision of risk ratings across insurers, and those with less accurate risk-rating algorithms tend to have more efficient cost structures. We assess the equilibrium effects of giving firms equal access to aggregated risk information from a centralized bureau. This policy significantly reduces prices by increasing competition, leading to a 13% boost in consumer surplus, nearly reaching the efficiency benchmark where firms have full knowledge of consumers' true risk. Aggregating information through the bureau favors low-risk consumers and reduces average costs by 18 euros per contract through more efficient insurer-insuree matching.
How can insurance design mitigate risk as opposed to just claims? We use novel sensor data to track drivers’ handheld phone use behavior (“HPU”), a generalization of text-and-drive in the smartphone age, and demonstrate how a behavior-based insurance contract can prevent accidents. Based on a prior experiment conducted, we develop a structural model to distinguish inattention, risk aversion, and the price elasticity of HPU. This facilitates counterfactual simulations of the “first-best” auto insurance contract with full insurance and a direct price on HPU (Holmstrom 1979). The contract features a 40-cent average charge per mile of HPU. A 62-cent charge can resolve additional externalities on traffic congestion and injuries.
(corresp. author | w/ Thomas Yu)
DIGITAL PLATFORM
AI Training for Online Entrepreneurs
We conducted a year-long experiment among millions of new sellers on an e-commerce platform. The treatment group received access to a free and comprehensive online entrepreneur training program. With a 24% take-up rate, training raised seller revenue (1.7% ITT and 6.6% ATT) largely by inducing advertising upfront. This allowed sellers to purchase sampling and overcome the ``cold-start'' problem. Our results reveal the information friction on---and the signaling value of---advertising for entrepreneurs on increasingly congested digital platforms.
Millions of businesses join advertising-supported digital platforms each year, yet their odds of success and how advertising affects entry are not well documented. Analyzing data from a major e-commerce platform in China, we find that fewer than 1% of newly registered and ID-verified sellers became small-to-medium-sized businesses. Records from consumer search sessions show that consumers are more likely to buy from entrants conditional on search, and that increasing exposure to entrants leads to more purchases, indicating an informational entry barrier. Finally, an experiment nudging some new sellers to advertise raised their traffic and revenues but also the rate of product returns, highlighting a trade-off between advertising’s ability to promote entry and its screening capabilities.
Many large digital platforms have begun to commercialize the enormous amount of data they have collected, both a byproduct of their expansion over the last two decades and a critical engine for their continued growth. We obtain data from a large e-commerce platform to study the efficacy of such a market-based solution for platforms' data and its implications on seller revenues and concentration.
Discussions and Other Writings
Discussion of Cosconati (2021):
The Effect of Insurance Telematics and Financial Penalties on Market-wide Moral Hazard
Telematics Data in U.S. Auto Insurance: Evidence from Telematics Contracts and an Accident Prevention Program
Discussion of "Data, Privacy Laws, and Firm Production: Evidence from GDPR" by Demirer, Jimenez Hernandez, Li, and Peng (2023)