Working Papers

Predicting choice averse and choice loving behaviors in a field experiment with actual shoppers” Forthcoming at Journal of Economic Behavior and Organization

A large body of chiefly laboratory research has attempted to demonstrate that people can exhibit choice-averse behavior from cognitive overload when faced with many options. However, meta-analyses of these studies, which are generally of one or two product lines, reveal conflicting results. Findings of choice-averse behavior are balanced by findings of choice-loving behavior. Unexplored is the possibility that many consumers may purchase to reveal their tastes for unfamiliar products, rather than attempt to forecast their tastes before purchase. I model such ‘sampling-search’ behavior and predict that the purchases of unfamiliar consumers increase with the available number of varieties for popular/mainstream product lines and decrease for niche product lines. To test these predictions, I develop a measure of popularity based on a survey of 1,440 shoppers for their preferences over 24 product lines with 339 varieties at a large supermarket in China. 35,694 shoppers were video recorded after the varieties they faced on shelves were randomly reduced. As found in the meta-studies, choice-averse behavior was balanced by choice-loving behavior. However, as predicted, the probability of choice-loving behavior increases with the number of available varieties for popular product lines, whereas choice-averse behavior increases with available varieties for niche product lines. These findings suggest that increasing the number of varieties has predictable opposing effects on sales, depending upon the popularity of the product line, and opens the possibility of reconciling apparently conflicting prior results.

Marrying up: Trading off spousal income and height,” (with Pierre-Andre Chiappori, Yu Yang, Junsen Zhang). Slides from the 2020 International Symposium on Contemporary Labor Economics

Couple’s heights tend to match. However, whether such matching is for the sake of height or the many desirable traits associated with stature (e.g., income) is unclear. We contribute novel experimental and empirical data to identify heterogeneity in preference for mate-height. We recorded clicks on profiles with randomly assigned height and income information on a major online dating website. These clicks reveal that taller men prefer taller women. By contrast, women not only prefer taller men but also higher income men, permitting the calculation of their marginal rates of substitution (MRS) for mate-height and income. Surprisingly, short women have the highest MRS for mate-height. We confirm this heterogeneity in preference for mate-height by applying the method of Chiappori et al. (2017) for multidimensional matching to data on married couples. Short early mothers drive these results. Our evidence is consistent with short women matching non-assortatively to increase the height of their children.

The College Admissions Contribution to the Labor Market Beauty Premium,” (with Man Xie and Junsen Zhang). Video, Slides from presentation at CES Session of ASSA, 2020

Beautiful people earn more. Surprisingly, this premium is larger for men than for women and is independent of the degree of customer contact. Overlooked is the possibility that beauty can influence college admissions. We explore this academic contributor to the labor market beauty earnings premium by sampling 1,800 social media profiles of students from universities ranked from 1 to 200 in China and the US. Chinese universities use only standardized test scores for admissions. In contrast, US universities use also grades and extracurricular activities, which are not necessarily beauty-blind. Consistent with beauty-blind admissions, student’s beauty is uncorrelated with the rank of their college in China. In the US, White men from higher ranked colleges are better-looking. As expected, the correlation is insignificant for White men who attended tech colleges and is highest for those who attended private colleges. We also find that White women and minorities of either gender are not better-looking at higher ranked colleges. Our evidence indicates a college admissions contribution to the labor market beauty premium for US White men, but not for students in China of either gender, White women, or minorities of either gender in the US, or for White men who attended technology colleges.

The Gender Difference in Mixed Strategy Nash Equilibrium Play” (with Gahye (Rosalyn) Jeon).

The mixed strategy Nash equilibrium (MNE) is a well-established concept in theory, but empirical and experimental support has been hampered by the need for a large sample size for across-subject choices and serial correlation for within-subject choices. We overcome these difficulties in testing for MNE play by exploiting a previously developed comparative statics effect for a common value pairwise all-pay auction which, strikingly, predicts that bids increase on perceived opponent’s risk tolerance, but not on the bidder’s own. Men bid as predicted, but women’s bids respond neither to their own nor to the opponent’s risk tolerance. Additionally, while the significance of men’s response to their beliefs about their opponent’s risk tolerance increases with the absence of mistakes on a pre-experiment quiz, women’s lack of response is unaffected by their quiz performance. These results are consistent with a prior finding of a gender difference in MNE play with a large sample of professional tennis players. We contribute by showing that MNE play can be tested with an ordinary sample of subjects in a laboratory setting. We show a gender difference in a factor that is likely crucial in most competitive situations: perceptions of the opponent’s risk tolerance.