Manuscripts under review and revision


  1. Sourcing under Supply Disruption and Responsibility Violation Risks: A Behavioral Investigation. Joint work with Karen Donohue and Karthik Natarajan. Reject and Resubmit at Management Science.
Summary (click to expand) We investigate sourcing decisions when faced with two suppliers with different cost and risk profiles. Sourcing from one supplier costs more but involves no risk, while sourcing from the other may introduce either supply disruption risk, which influences product supply, or responsibility violation risk, which influences customer demand. To contrast these two types of supplier-induced risks, we present a parsimonious comparative framework based on risk matrices used in practice that organize risk level by two dimensions: likelihood and impact. We first analytically characterize the profit-maximizing sourcing strategies in different likelihood/impact environments and then employ an incentivized experiment with human participants to explore actual sourcing behavior. While sole-sourcing is always theoretically optimal, we find that individuals tend to diversify. However, they are less likely to diversify and more likely to select the optimal supplier when faced with responsibility violation risk relative to supply disruption risk. Concerning the underlying environment, we find that buyers respond similarly to changes in risk likelihood and impact levels in terms of the adjustment in order allocation between the two suppliers. This behavior deviates from the normative predictions, and our analysis suggests that managers will benefit from recognizing the unique characteristics of the two risk dimensions and developing tailored strategies to manage each type of risk. The negative influence of these behavioral factors on profit can be quite significant, especially under responsibility violation risk.

  1. Pricing Online Delivery Options: Aggregated vs. Disaggregated Customer Rating. Joint work with Tim Huh. Under review at Manufacturing & Service Operations Management.
Summary (click to expand) Online shopping has become an indispensable part of people's lives. Online retailers, especially small ones who sell through established e-commerce platforms, often choose to outsource the delivery functionality to a set of delivery service providers. To help resolve uncertainty associated with these delivery service providers' on-time reliability, customers can rate the reliability of these providers. A retailer may choose to adopt either an aggregated rating system under which only the overall rating across all the delivery service providers is shown or a disaggregated rating system under which each provider's rating is displayed separately. Online retailers need to understand how to set optimal prices for different delivery options and recognize potential implications of rating system design. Our analysis shows that under the aggregated rating system, the optimal markups should be differentiated, while under the disaggregated rating system, every option should have the same markup. However, the retailer's optimal expected profit is equivalent across the two rating systems. Hence, the simpler aggregated rating system may achieve the same expected outcomes as the seemingly more informative disaggregated rating system. We also assess how the rating systems impact delivery service providers and customers.

Work in progress


  1. Vaccine Nationalism: Implications of Government Mandates on Vaccine Manufacturing and Public Health Outcomes. Joint work with Karen Donohue and Karthik Natarajan. Analysis in progress.
Summary (click to expand) The COVID-19 pandemic has created unprecedented pressure on healthcare supply chains. Due to the limited supply of critical products such as PPEs and drugs, several national governments have implemented policies to limit the exporting of such products. As the success of combating the pandemic hinges on the availability of vaccines, there are growing concerns that governments might take similar measures to prioritize supplying vaccines to their own citizens, a phenomenon dubbed by the media as "vaccine nationalism." Against this backdrop, we investigate potential implications of such government mandates on vaccine manufacturers' capacity investment decisions and the associated public health outcomes. Our analysis shows that these mandates do not necessarily result in the intended consequences of increasing the local availability of vaccine doses. In some situations, the policies may actually induce the manufacturer to diversify production to another country. Also, the local, and sometimes even total, capacities may decrease if the capacity investment cost is high. We further study how the associated changes in capacity investments may influence public health outcomes and what behavioral factors may influence decision-makers' choices.

  1. Individual Unlearning in Process Management. Joint work with Kevin Linderman and Brent Moritz. Pilot studies completed.
Summary (click to expand) Process management requires organizations to learn and adapt to new routines. While prior research primarily adopts an organizational learning perspective to study process and knowledge management, the "unlearning" procedure attracts more and more attention in recent years. This intentional effort in discarding old knowledge in order to embrace new knowledge is even considered a source of dynamic capability that organizations can leverage on. Although prior literature on organizational learning and change management has recognized individuals' role in shaping organizational level outcomes, there is relatively little research on how individual unlearning might contribute to firms' process improvement initiatives. Our study adopts an experimental approach to address this gap and develop a better understanding of the microfoundations of organizational unlearning in process management. We examine what kind of informational cues might help promote the unlearning process and improve learning outcomes and explore whether important individual characteristics, such as personality, influence different cues' efficacy. Preliminary data shows that different individuals may be susceptible to different cues and that there might be potential benefits of tailoring different cues to individuals with different characteristics.