UC Berkeley

In my research, I use modern industrial organization theory and econometrics to study the impact of behavioral frictions among buyers and sellers in different market environments on market outcomes and welfare. Following the seminal works of Becker (1957) and Arrow (1972), it is common knowledge that market competition mitigates such non-cognitive frictions as employer taste-based prejudice. Hence, it is not surprising that today, almost half a century after Kahneman and Tversky (1979) introduced their Prospect Theory, many economists believe that market forces —the invisible hands of sophisticated profit-maximizing suppliers—mitigate supply-side cognitive-based frictions as well. In my research, I challenge these popular views and show that market forces do not necessarily compete away the effects of behavioral cognitive or non-cognitive frictions in firm pricing and employment.

The dissertation consists of three chapters: (i) "Behavioral Professionals: Evidence From the Commercial Auto Insurance Industry," (ii) "Price and Prejudice: Customer Taste-Based Discrimination and Competition," and (iii) "The Lemons Gap: Demand For Insurance of Quality Uncertain Goods."

In Chapter 1, I present my work "Behavioral Professionals: Evidence From the Commercial Auto Insurance Industry." A cornerstone of the IO study of selection markets is that competition disciplines sellers to customize coverage and premiums optimally. But is this the case? Using data from one of the largest Israeli commercial auto insurance providers, an affiliate of a multinational insurance company, I find there is too little adjustment in the intensive margin. Premiums barely change with expected costs as projected by pre-determined factors (vehicle age) and signals (claim history). At the same time, I find there is too much adjustment in the extensive margin, with an excessive denial of insurance in response to recent claims. Using unique grading documents, I integrate the insurer’s subjective risk assessment into the study of insurance markets. I find that the insurer's risk assessment outweighs recent claims and misevaluates vehicle age. Structural model estimates suggest that insurers enjoy incumbency advantages over their own customers, and clients are rationally inattentive to competitors’ pricing unless they are faced with a price increase. Both channels allow sub-optimal behavior to persist. Finally, I find that supply-side behavioral frictions, which result in excessive denial, mainly harm disadvantaged customers - single-fleet clients of old vehicles\textemdash and diminish with the client's fleet size.

In Chapter 2, I present my work "Price and Prejudice: Customer Taste-Based Discrimination and Competition." This work investigates the effect of competition on the incidence of tastes for discrimination. The model shows that monopolistic sellers discipline discriminatory buyers by taxing their taste for discrimination. In equilibrium, monopolistic sellers hire a lower share of White workers and pay them a lower premium than sellers in a competitive market. These results are tested in the context of the US banking deregulation that affected product market competition, and I quantify its impact on customer-driven labor market discrimination. Using Census/ACS data from 1960 to 2010, O*NET measures of job requirements, and GSS measures of discriminatory attitudes against Black by state, I find that the Black-White wage and employment gaps increased following bank deregulation in jobs requiring intensive contact with clients, especially in states with high measures of prejudicial preferences.

In Chapter 3, I present my work "The Lemons Gap: Demand For Insurance of Quality Uncertain Goods." This work studies the difference between insuring a quality uncertain good and a monetary loss. I integrate key insights from the pre-owned market into the analysis of the demand for insurance. I find that adverse selection in the resale market results in a missing insurance market. There's a gap between the insured vehicle and the resale market’s quality, especially for new vehicles. As a result, clients over-insure their quality uncertain goods, yet demand drops over the vehicle life cycle. The partial compensation further amplifies over-insurance patterns driven by behavioral attributes. The gap results in time trends. As the vehicle ages, demand drops, the insurance market is more adversely selected, and moral hazard increases. The incomplete compensation can result in context-dependence demand for insurance, customers' over-insurance limited risk, in general, and for durable goods, in particular.