Astorne-Figari, C. and Yankelevich, A. Consumer Search with Asymmetric Price Sampling. Economics Letters 122/2 (2014), pp. 331-333. (Extended version here)
Are Changes of Major Major Changes? The Roles of Grades, Gender, and Preferences in College Major Switching (with Jamin Speer )
The choice of college major is a key stage in the career search. Almost half of college graduates switch majors at least once, suggesting that major choice is a process rather than a single decision. This paper provides the first comprehensive analysis of major switching, modeling the choice process as a matching problem between student and major on both academics and preferences. Low grades signal academic mismatch and predict switching majors - and the lower the grades, the larger the switch in terms of course content. Additionally, when students switch majors, they switch to majors that “look like them”: females to female-heavy majors, blacks to black-heavy majors, and so on. Lower-ability women flee competitive majors at high rates, while higher-ability women and men of all abilities are undeterred. Women are far more likely to leave STEM fields and to switch to majors that are less competitive – but not less science or math-intensive – suggesting that leaving STEM is more about fleeing the “culture” and makeup of STEM majors than it is about fleeing science and math.
Men and women respond differently to early-college struggles. Men are more likely than women to drop out of college, while women are more likely to switch majors. These effects offset so that there is no gender gap in the probability of graduating in one’s initial major choice. For students who begin in STEM majors, however, women are far less likely to graduate in the field, driven bythe fact that they are twice as likely to switch majors. We find no evidence that women are more sensitive to poor academic performance in the switching or dropout decisions.
Advertising and Pricing When Attention is Limited (with Joaquin Lopez and Aleksandr Yankelevich)
We analyze markets where firms that compete on price advertise to vie for consumers' limited attention. Our baseline model of attention has a number of desirable properties. It offers an explanation for price dispersion in homogeneous goods markets in the absence of search costs. Moreover, it leads to a unique symmetric price distribution that changes from competitive to monopoly pricing as attention becomes more limited. When firms can influence consumer attention by advertising and the cost of advertising is low, advertising leads firms to a prisoner’s dilemma that adversely impacts profits without changing prices. However, moderately costly advertising permits firms to raise prices and possibly profits by segmenting the market.
An empirical investigation on the transfer of expatriates within MNCs from a knowledge perspective (with Joonhyung Lee )
We study the use of expatriates in transferring tacit knowledge within a multinational corporation (MNC). Using data from MNCs headquartered in South Korea, we exploit the fact that South Korea is significantly different from other countries in terms of language, culture, and ethnic diversity to argue that unobservable knowledge can be traced by observable labor mobility. We use a model of knowledge hierarchies applied to MNCs in order to generate testable predictions about the optimal allocation of knowledge between headquarters and its foreign affiliates. Our results are consistent with these predictions. First, the use of expatriates in a foreign affiliate increases as communication between South Korea and the host country becomes more costly. Second, the use of expatriates is negatively correlated with the scale of the foreign affiliate. Additionally, the extent to which the use of expatriates relates with communication costs and scale depends on the sectoral complexity.
Expectations and Follow-Through: The Roles of Confidence and Non-Cognitive Skills for Self-Employment (with Andrew Hussey and Daniel Mangrum)
We examine the predictors of both long-term expectations of self-employment and future self-employment activities and earnings among the same individuals, with a particular focus on gender differences and the roles of non-cognitive skills. Using longitudinal data from the GMAT Registrant Survey, which includes prospective graduate management students, our analysis involves wide-ranging and novel sets of variables, including work-life balance and job preferences, self-efficacy, confidence, and other non-cognitive skills or characteristics. We find notable differences in the drivers of self-employment and self-employment expectations between men and women, and also large differences in the set of variables that relate to self-employment intentions versus future self-employment outcomes. While preferences for work-life balance matter more for men's expectations, preferences about non-monetary characteristics of the job, such as job security and interesting work, matter more for women. In contrast, regarding actual self-employment, only non-cognitive skills play a substantial role for women, while men are driven mostly by preferences over work-life balance. Confidence in one's quantitative skills influences self- employment decisions, especially for women, and it also affects success in both the self-employed and the traditionally employed sectors, as reflected in earnings.
Kin Targeted Altruism with Noise
Can pure altruism generate strategic altruism when kin recognition is noisy? This paper studies a prisoners' dilemma played between two people who exhibit altruistic preferences towards kin. The probability that a player's opponent is kin is common knowledge. Instead of observing the degree of relatedness with the other player, each player observes a noisy private signal. When the game is played once, players cooperate only with those identified as kin. However, when the prisoners' dilemma is played for two periods instead of one, uncertainty about relatedness brings strategic considerations into the game even if the odds of being related are small. De- pending on the parameters, there are Perfect Bayesian Equilibria in which players cooperate in the first round even after getting a negative kin signal. Since a player can make inferences about her opponent’s signal based on first period actions, a non-relative mimics kin to induce cooperation in the second period.