Selected Publications
Targeted Advertising and Consumer Inference
( Shin, Jiwoong | YU, Jungju )MARKETING SCIENCE2021-09
Abstract
The mere fact that consumers are targeted by advertisements can affect their inference about the expected utility of a product. We build a micro-model where multiple firms compete through targeted advertising. Consumers make inferences from targeted advertising about their potential match values for the product category, as well as the advertising firm's unobserved quality. We show that in equilibrium, upon being targeted by a firm, consumers make optimistic inferences about the product category and the firm's quality. With such improved beliefs, a targeted consumer is more likely to engage in a costly search throughout the category. We find that the increase in consumer search creates an advertising spillover beyond the level of the mere awareness effects of advertising and that firms' equilibrium level of targeted advertising can be non-monotonic in targeting accuracy. Additionally, we show that sometimes, it can be optimal for firms to relinquish customer data and instead engage in non-targeted advertising. The results provide insights into the trade-offs between advertising reach and targeting accuracy.
Simple Mechanisms for Sequential Capacity Allocations
( Kim, Bosung | Kim, Sang Won | Iravani, Seyed M. R. | Park, Kun Soo )PRODUCTION AND OPERATIONS MANAGEMENT2021-09
Abstract
We consider a sequential capacity allocation problem of a firm to its retail branches that sell the firm's product. The orders of the retail branches to the headquarters arrive sequentially, and each allocation decision has to be made before the next order arrives. The objective of the headquarters is to maximize the overall firm profit, that is, the total profit of all the retail branches. Each retail branch makes ordering decision independently by using private information about its local market condition in order to maximize its own profit. Hence, they may strategically inflate their order quantities in their favor, potentially hurting the firm profit. We first discuss the importance of capacity rationing in maximizing the firm's profit by finding the first-best allocation outcome, the optimal solution without information asymmetry. Based on this, we design mechanisms that effectively overcome the information asymmetry. First, we design a simple threshold-type mechanism where truthful reporting is optimal and capacity rationing is implemented by limiting allocation beyond a pre-specified threshold. We show this mechanism is optimal within the class of mechanisms that do not allow any side payments. We also design a mechanism with side payments that is optimal among all possible mechanisms. In particular, we show that this payment-based mechanism achieves the first-best allocation, fully overcoming information asymmetry. Although optimal, it may not be practical because of the complexity in the side payment menu, so we also propose a simple variant of it with only a few parameters. Our extensive numerical study shows that our simple threshold-type and payment-based mechanisms achieve near-optimal performance.
Drinking Through Good Times and Bad: The Role of Consumer Differences
( Joo, Hailey Hayeon | Kim, Minki | Lee, Jungmin | Chintagunta, Pradeep )JOURNAL OF MARKETING RESEARCH2021-08
Abstract
The authors revisit the question of alcohol consumption and public health over business cycles by decomposing overall alcohol consumption into drinking frequency and intensity in relation to consumer heterogeneity. To study this question, they use consumer-level panel data on the reported consumption (not purchases) of beer, which is the most heavily consumed alcoholic beverage and accounts for the majority of binge drinking in the United States. Leveraging the panel nature of the data, the authors find a negative (positive) relationship between unemployment and drinking frequency (intensity). Total consumption, which is the product of drinking frequency and intensity, is procyclical. To uncover differences in behavior across consumers and to provide policy recommendations at a segment level, the authors present a structural model where consumers simultaneously choose the frequency and intensity of their alcohol consumption. They find differences across consumers in their behaviors, notably with respect to income and age. They conduct policy simulations to compare the effectiveness of alcohol-related policies to counter the adverse effects of recessions on the health of vulnerable groups such as low-income and elderly populations.
Arbitrage Portfolios
( Kim, Soohun | Korajczyk, Robert A | Neuhierl, Andreas )REVIEW OF FINANCIAL STUDIES2021-06
Abstract
We propose a new methodology for forming arbitrage portfolios that utilizes the information contained in firm characteristics for both abnormal returns and factor loadings. The methodology gives maximal weight to risk-based interpretations of characteristics’ predictive power before any attribution is made to abnormal returns. We apply the methodology to simulated economies and to a large panel of U.S. stock returns. The methodology works well in our simulation and when applied to stocks. Empirically, we find the arbitrage portfolio has (statistically and economically) significant alphas relative to several popular asset pricing models and annualized Sharpe ratios ranging from 1.31 to 1.66.
The Deterrent Effect of Ride-Sharing on Sexual Assault and Investigation of Situational Contingencies
( Park, Jiyong | Pang, Min-Seok | Kim, Junetae | Lee, Byungtae )INFORMATION SYSTEMS RESEARCH2021-06
Abstract
Sexual assault is one of the most repellant and costly crimes, which inflicts irrecoverable harms on victims and society. This study examines the effect of information technology (IT)-enabled ride-sharing platforms on sexual assaults. Drawing upon routine activity theory from the criminology literature, we posit that ride-sharing can reduce a passenger's risk of being a suitable target of sexual assault by providing a more reliable and timely transportation option for traveling to a safer place. By exploiting the nationwide quasi-experimental setting of Uber's city-by-city roilouts in the United States during 2005-2017, we demonstrate that Uber's entry into a city is negatively associated with the number of rape incidents. To zoom into the effects of ride-sharing at a more granular level, we employ precinct-hour-level data on Uber pickups and rape occurrences in New York City in 2015 and conduct spatiotemporal analyses. Our results from the spatiotemporal analyses corroborate those of the quasi-experiment and further reveal situational contingencies in the deterrent effect of ride-sharing. Specifically, ride-sharing contributes to a more significant reduction in the likelihood of rape occurrences in neighborhoods with limited transportation accessibility, and ride-sharing is more effective in deterring sexual crime in riskier circumstances, such as around alcohol-serving places on weekend nights or when the probability of crime occurrences increases. This study sheds new light on the potential of IT-enabled platforms to improve social well-being beyond their economic contributions and offers a new theoretical insight on the distinct role of digital platforms in public safety.
