UC San Diego

This dissertation contains three essays studying topics in applied microeconomics. The first chapter studies the formation and the spread of crisis-driven racial animus during the coronavirus pandemic. Exploiting plausibly exogenous variation in the timing of the first COVID-19 diagnosis across US areas, we find that the first local case leads to an immediate increase in local anti-Asian animus, as measured by Google searches and Twitter posts that include a commonly used derogatory racial epithet. This rise in animus specifically targets Asians and mainly comes from users who use the epithet for the first time. These first-time ch-word users are more likely to have expressed animosity against non-Asian minorities in the past, and their interaction with other anti-Asian individuals predicts the timing of their first ch-word tweets. Moreover, online animosity and offline hate incidents against Asians both increase with the salience of the connection between China and COVID-19; while the increase in racial animus is not associated with the local economic impact of the pandemic. Finally, the pandemic-driven racial animus we documented may persist beyond the duration of the pandemic, as most racist tweets do not explicitly mention the virus.

The second chapter investigate if primary care physician (physician henceforth) and patient concordance in terms of socio-economic status (SES) reduces the SES inequality in health. We exploit variations in SES concordance between physicians and patients that are induced by plausibly exogenous clinic closures. We find that SES concordance lowers low-SES patients' mortality while high-SES patients' mortality does not depend on their physicians' SES. Together, these effects translate to a 23% reduction in the SES-mortality gradient. Mortality reductions related to cardiovascular conditions are especially pronounced. We study patients' health behavior and physicians' treatment choices to explain how SES concordance reduces patient mortality. Low-SES patients with low-SES physicians receive more care at the intensive margin; making more office visits per year and receiving more services per visit. In addition, they are more likely to be prescribed Statins, adhere to diabetes check-up visits, and are less likely to have avoidable hospitalizations due to COPD, relative to comparison groups.

The third chapter asks: how does employer reputation affect the online labor market? We investigate this question using a novel dataset combining reviews from Glassdoor.com and job applications data from Dice.com. Labor market institutions such as Glassdoor.com crowd-sources information about employers to alleviate information problems faced by workers when choosing an employer. Raw crowd-sourced employer ratings are rounded when displayed to job seekers. By exploiting the rounding threshold, we identify the causal impact of Glassdoor ratings using a regression discontinuity framework. We document effects from both labor demand and supply sides at equilibrium. We find that displayed employer reputation affects employer’s ability to attract workers, especially when the displayed rating is sticky. Employers respond to the rounding threshold by posting more new positions and re-activating more job postings. The effects are the strongest for firms that are private, smaller, and less established, suggesting that online reputation is a substitute for other types of reputation.