UC Berkeley

My dissertation consists of three studies that incorporate behavioral economics element in analyzing decisions of individuals in the labor and housing market.

The first chapter studies how college students learn about the earning opportunities associated with different majors. I use data from two major longitudinal surveys to develop and estimate a learning model in which students update their expectations based on the contemporaneous earning realizations of older siblings and parents. Reduced-form models show that the probability of choosing a major that corresponds to the occupation of an older sibling or parent is strongly affected by whether the family member is experiencing a positive or negative earnings shock at the time the major choice is made. Building on this finding, I estimate a model of major choice that incorporates learning from family-based information sources. The results imply that students overestimate the predictive power of family members' earnings: the decision weight placed on family wage realizations is much larger than can be justified by the empirical correlation between their own earnings and their family members' earnings.

My second chapter focuses on how time preference affect their job searching under unemployment insurance (UI) policies. Previous studies find that higher UI benefit, extended UI eligibility duration, bonus payment or severance pay affects unemployed workers' job-finding hazard rate but not the subsequent job match quality. I construct and estimate a dynamic job search model endogenizing both the search intensity and reservation wage with hyperbolic discounting. Using data from several state job bonus experiments from the 1980s (the Illinois UI Incentive Experiments), I find the model with hyperbolic discounting fits the effect of the job-bonus treatment better, and an unemployed worker's reservation wage decreases slower during search duration under the hyperbolic discounting framework, implying that bonus payments induce higher search effort but do not significantly decrease workers' reservation wages.

The third chapter is a joint work with Tristan Gagnon-Bartsch and Antonio Rosato. In this study, we propose and empirically test a theoretical model of loss aversion in the housing market. Compared to the empirical findings of Genesove and Meyer (2001), our model makes a new prediction: sellers who suffer a relatively small loss (when the current market value is lower than the previous purchasing price) will set prices equal to their original purchase price. Hence the model predicts an asymmetric distribution of gains to sellers which assigns less mass to small negative values than to equally size positive values and has a spike at zero. We first use the same data-set used by Genesove and Meyer to test our new prediction and find that between 4% and 10% sellers incurring a loss ``bunch'' by asking a price within 5,000 dollars of the original purchasing price. We also collect new real-estate data from the San Francisco Bay Area in 2011 and find that the pricing behavior of individual sellers is still consistent with loss aversion.