Research

As the biggest US presidential election upset in recent decades, Donald Trump’s 2016 victory introduced new uncertainties in the trade environment, including the potential for additional prohibitive tariffs. While it is well-documented in the trade literature that firms anticipate and respond to scheduled tariff changes, it remains unclear whether this still holds when changes are uncertain and may happen at any moment, for example, following a contentious election. Focusing on the period between the 2016 election and the first round of China-specific tariffs in mid-2018, I empirically study if and how American importers reacted to potential tax hikes using US Customs bills of lading data. By exploiting cross-product, origin, and time variations in tariff risks through a triple-difference approach, I find that importers facing higher risks stockpiled in response to Trump’s election—increasing their quarterly imports by around 5% through larger order sizes rather than increased order frequency. However, there is no evidence that firms expanded their trade network, diversified their sourcing portfolio, or diverted away from China during this period. The stockpiling behavior was more pronounced among smaller importers, which can partially be explained by the downstream nature of their purchases rather than differences in post-election entry/exit rates or input storability.

Transportation in international trade often involves transshipment by entrepôt hubs as part of the global hub-and-spoke shipping network. We investigate the returns from being a hub country—the impact of global transshipment activity by these hubs on their own international trade flows and supply chains over a 10-year period. We establish that the majority of US imports–especially those from smaller origin countries—are transshipped via a select few hub countries, and transshipment activity is positively correlated with the hub country’s total trade, particularly through its exports. We then unpack how transshipment activity contributes to exports. Exploiting the indirect nature of trade, we find that transshipment activity increases the hub country’s imports from the origin country and its exports of downstream goods. This novel trade channel from transshipment activity highlights the important link between global value chains and transportation, and a potential venue for developing countries to further participate in global value chains.

Contract is the quintessential aspect of business transactions and trade between firms. This project delves into one key aspect of all contracts—duration—and asks whether the length of a contract affects a firm’s decisions to adopt new technologies and spend more on R&D, particularly in the case of a developing nation such as Vietnam. Theoretically, the effect of contract length on investments should be increasing at a decreasing rate: short-term contracts generate uncertainty, additional search costs, and a lack of financial stability that is required for investments; on the other hand, very long-term contracts, despite offering greater stability, risk the principal-agent problem in which the contracted firm slacks off. Additionally, this project studies the kinds of technology that a firm chooses to adopt, if it is induced to do so by a longer contract.

I model the tariff negotiating process between two large countries as a Rubinstein's alternating-offers bargaining game with outside options. The outside options are allowed to be dependent on country strengths, i.e., their respective level of influence over the rest of the world, instead of being some arbitrary static numbers. The model has a unique subgame perfect equilibrium, and comparative statics show that the equilibrium tariff that a country sets is increasing in its own strength and decreasing in its opponent's strength. Equivalently, a country should be able to set a higher tariff and enjoy a larger slice of the welfare surplus cake when it has a better outside option compared to its opponent. Using data on trade protection from the Market Access Map database, I test these predictions and find mixed empirical evidence of the theory.

Cram schools, also known as “extra classes,” are a widespread phenomenon across many East and Southeast Asian countries, including Vietnam. Using the Vietnamese national survey data, this project studies the determinants of household demand for extra classes and whether more spending on cram schools translates to better employment opportunities for young college graduates. In doing so, this project strives to weigh the costs and benefits of sending kids to cram schools for a household, thus helping educational policy makers better understand and regulate this market.

The Comprehensive Osteopathic Medical Licensing Examination of the United States (COMLEX-USA) is a three-level, national standardized osteopathic medical licensing examination administered by the National Board of Osteopathic Medical Examiners (NBOME). Using data from a selective osteopathic medical school, we examine the predictive power of key academic metrics on the probability that a student will pass the COMLEX-USA Level 1 exam on their first attempt. We also calculate the optimal COMSAE cut score which students are required to reach before they can take the COMLEX-USA Level 1 to maximize their chance of passing.

The COMLEX-USA Level 2-CE is the second exam in a series of three osteopathic medical licensing examinations, which is typically taken by osteopathic medical students at the end of their third year and considered an important determinant of profile competitiveness in the match process. To help academic advisors and osteopathic medical students choose their application strategy, we develop a process to predict the score distribution for each student using Bayesian modeling. Notably, our process allows for: (1) recurrent data updates as new test scores come in over time, (2) iterative imputation of missing data, and (3) periodic production of reproducible individualized reports.

Up to a third of companion animals receive nutraceuticals/supplements for a variety of conditions, including cancer. Cancer-targeted supplements generally rely on antioxidant properties. Since several chemotherapy drugs produce free radicals as part of their cytotoxic mechanism, the potential for antagonism with supplements has been a long-standing concern but has not been formally studied. As the first step to study this issue, we circulated an online survey to clients who presented a canine patient to the oncology service at a Veterinary Medical Center between 2016 and 2022 to gauge the extent and type of supplements used. Our findings indicate that: (i) Over half of the surveyed sample provided supplements to their dogs with various tumor types; (ii) The most common antioxidant supplements used were omega 3 fatty acids, cannabidiol (CBD)/hemp products, turkey tail mushroom, and canine multivitamins; (iii) There was no statistically significant association between tumor type and the likelihood of antioxidant supplementation, and there was no significant association between antioxidant use and client election for cytotoxic therapy. These findings further support the need for additional investigation regarding potential interaction between antioxidants and various cancer treatments.