Supply Chain & Operations Events

The Supply Chain and Operations Department & Juran Research Center Seminar Series showcases current research in Management and Operations Science, including topics in operations strategy, technology, quality, new product development, and supply chain management. Speakers are drawn from universities around the world.


2023-2024 Spring

Jonathan Eugene Helm

Kelley School of Business, Indiana University

Date: Friday, February 9, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

Abstract:
The opioid crisis has ravaged the United States, taking 69,000 lives in 2020, with prescription opioids accounting for 98% of opioid abuse. Though this epidemic is often considered a White public health crisis nationally, overdose deaths have doubled among people of color from 2017-2019. Significant public criticism has gone to several individual firms involved as research has demonstrated that the opioid crisis was driven by excessive supply from the pharmaceutical industry. We find evidence that the scope of the blame goes beyond individual actors to include the very structure of supply chains, where complex supply chains may have exacerbated the crisis by dispensing significantly more opioids. We posit that supply chain complexity allowed mass quantities of opioids to escape detection by the Drug Enforcement Administration (DEA). Further, we find new evidence showing the greater impact of complexity on dispensing in non-White communities, which underscores their exclusion from the public discourse and governmental response surrounding the crisis and suggests possible racial bias in the DEA’s regulatory policies. Our analysis was made possible by the hotly contested 2019 release of the DEA’s ARCOS database that logs every shipment in the supply chain of opioids in the United States from 2006-2014. Using a fixed effects model, we find that a one-unit increase across three dimensions of supply chain complexity is associated with a 16% increase in opioid dispensing. This effect is intensified in non-White communities, where a 10% increase in the non-White population is associated with a 3.39% (1.33%) increase in the effects of complexity for high (average) complexity supply chains. To verify that excess dispensing of high-complexity pharmacies supplied non-medical/recreational demand, we exploit the reformulation of OxyContin designed to prevent recreational use as an exogenous shock to the market. In a novel approach, we use the fact that different pharmacies received their first shipment of reformulated OxyContin at different times and use a difference-in-differences model to estimate the heterogeneous effect of the shock on dispensing. As the abuse-deterrent OxyContin stifled demand, high complexity pharmacies experienced a 15.31% greater reduction in dispensing compared to lower-complexity pharmacies, suggesting the excess dispensing was satisfying non-medical/recreational demand.

 

Michelle Shell

The Tuck School of Business, Dartmouth

Date: Friday, April 5. 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132

 

Saed Alizamir

Darden School of Business, University of Virginia

Date: Friday, April 19, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

 

Telesilla Kotsi

Fisher College of Business, Ohio State

Date: Friday, May 10, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132

 

Ashish Kabra

Robert H. Smith School of Business, University of Maryland

Date: Friday, May 24, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132


2023-2024 Fall

 

Suvrat Dhanorkar

Smeal College of Business, Penn State

Date: Friday, September 22, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"Does Legalizing Marijuana Increase Toxic Waste?"

Evidence from Manufacturing Facilities in the United States"

Abstract:
Problem Definition: Marijuana is increasingly gaining legal acceptance in the United States. From 1996 to 2016, 28 states and the District of Columbia legalized marijuana for medical use. At the same time, studies find that when individuals use marijuana, they can experience several negative consequences such as reduced attention, hampered memory, and poor task completion—which implies that marijuana can degrade the output of human resources. But human resources are a critical input into manufacturing and are vital for ensuring operational effectiveness, which suggests that increased access to marijuana could adversely affect manufacturing operations. However, hardly any work has explored the impact of legalizing marijuana on manufacturing activities. In this study, we seek to bridge the gap by examining the effect of legalizing marijuana on the toxic chemical releases of manufacturing facilities.

Methodology/Results: We leverage a state-level quasi-experimental setting that evolves from the staggered enactment of marijuana legislation by different states in the U.S. We augment this setup with data from the Environmental Protection Agency’s Toxics Release Inventory program. Our dataset spans 1987-2016 and includes details on the toxic chemical releases of 45,720 manufacturing facilities in the U.S. We find that medical marijuana legislation (MML) adversely affects the toxic releases of facilities in the state—the average releases increased by 5.22% after MML. Further examination reveals that facilities undertake fewer managerial and technical modifications to their operational processes, which clarifies the mechanisms that affect the toxic releases. Finally, we find that recreational marijuana legislation (RML), which increases access to marijuana for the general population, leads to further increases of toxic releases—the effect goes beyond the impact of MML.

Managerial Implications: Our study provides important insights for managers and policy makers by casting light on the detrimental impact of legalizing marijuana on manufacturing operations.

 

Priyank Arora

Darla Moore School of Business, University of South Carolina

Date: Friday, October 6, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"Farm Equipment Sharing in Emerging Economies"

Abstract:
In emerging economies, a growing number of farm equipment sharing platforms have emerged to connect smallholder farmers with tractor owners who are willing to fulfill farmers’ requests for mechanization services. Due to the small farm sizes and the low digital literacy in rural areas of emerging economies, these platforms often rely on the so-called “booking agents” to collect demand from individual farmers and submit the aggregated demand on the platform. This is in contrast to other conventional sharing platforms where no booking agents are present and the service providers directly fulfill service requests from individual customers. In this paper, we explicitly capture the role of booking agents on a farm equipment sharing platform and study how the platform should choose the price and wage rates to appropriately incentivize all entities on the platform (including farmers, tractor owners, and booking agents). Although the farm equipment sharing platforms are popularly referred to as “Uber for Tractors,” our analysis offers insights about when and why these platforms should not rely on conventional wisdom derived from other conventional sharing settings (such as ride-sharing), due to the presence of booking agents. Further, inspired by the increasing efforts in practice from governments or donor agencies to enhance the supply and demand sides of these platforms (such as by increasing the number of tractors on the platform or by making booking agents’ demand aggregation more efficient), our analysis also sheds light on how such efforts affect the platforms’ optimal decisions and the equilibrium outcomes. Finally, we present a calibrated numerical study using data from Hello Tractor—an award-winning farm equipment sharing platform—to provide an illustration of how the generated insights can map to practice.

*Joint work with: Olufunke Adebola (Deloitte; previously, Hello Tractor) and Can Zhang (Fuqua School of Business, Duke University)

 

Meng Li

Bauer School of Business, University of Houston

Date: Friday, October 27, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"The Value of AI in OM: Evidence from Field Experiments"

Abstract:
I will discuss two papers. In the first paper “AI and Procurement,” we study how buyers’ use of artificial intelligence (AI) affects suppliers’ price quoting strategies. Specifically, we study the impact of automation—that is, the buyer uses a chatbot to automatically inquire about prices instead of asking in person—and the impact of smartness—that is, the buyer signals the use of a smart AI algorithm in selecting the supplier. We collaborate with a trading company to run a field experiment on an online platform in which we compare suppliers’ wholesale price quotes across female, male, and chatbot buyer types under AI and no recommendation conditions. We find that, when not equipped with a smart control, there is price discrimination against chatbot buyers who receive a higher wholesale price quote than human buyers. In fact, without smartness, automation alone receives the highest quoted wholesale price. However, signaling the use of a smart recommendation system can effectively reduce suppliers’ price quote for chatbot buyers. We also show that AI delivers the most value when buyers adopt automation and smartness simultaneously in procurement.

In the second paper “Physician Adoption of AI Assistant,” we study AI assistants---software agents that can perform tasks or services for individuals---which are among the most promising AI applications. However, little is known about the adoption of AI assistants by service providers (i.e., physicians) in a real-world healthcare setting. In this paper, we investigate the impact of AI smartness (i.e., whether the AI assistant is empowered by machine learning intelligence) and the impact of AI transparency (i.e., whether physicians are informed of the AI assistant). We collaborate with a leading healthcare platform to run a field experiment in which we compare physicians’ adoption behavior, i.e., adoption rate and adoption timing, of smart and automated AI assistants under transparent and non-transparent conditions. We find that smartness can increase the adoption rate and shorten the adoption timing, while transparency can only shorten the adoption timing. Moreover, the impact of AI transparency on the adoption rate is contingent on the smartness level of the AI assistant: the transparency increases the adoption rate only when the AI assistant is not equipped with smart algorithms and fails to do so when the AI assistant is smart. Our study can guide platforms in designing their AI strategies. Platforms should improve the smartness of AI assistant. If such an improvement is too costly, the platform should transparentize the AI assistant, especially when it is not smart.

 

Ken Moon

Wharton School, University Of Pennsylvania 

Date: Friday, November 3, 2023
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

 

"Bringing Data Science to the Management of Workforces"

Abstract:
The talk will cover several, real-world collaborations relating to the operational management of workforces.  The main part of the talk will focus on a research project with the Apple Worker Exit Study using extensive data on staffing, productivity, and pay from within a consumer electronics supply chain producing tens of billions in USD revenue quarterly.  We study how firms should manage the problem of worker turnover, including its surprising impact on low-skilled workforces and the implications for production, wage, and inventory decisions.  Despite the lack of skills, we find that worker turnover impedes coordination between assembly line coworkers by weakening knowledge sharing and relationships.  We structurally estimate a dynamic equilibrium model of workers’ endogenous turnover decisions and the firm’s dynamic production and staffing decisions, and we apply reinforcement learning to evaluate managerial alternatives.  A less turnover-prone, hence more productive, workforce reduces the firm’s variable production costs by 4.5%, or an estimated $928 million for the studied product. Such benefits justify paying higher efficiency wages even to less skilled workforces; furthermore, interestingly, rational inventory management policies incentivize self-interested firms to reduce rather than tolerate turnover.   We also cover more recent research that develops learning algorithms to address the problem of worker stress and burnout for highly skilled workforces (ICU nurses and fighter jet pilots).  In particular, we equip the nurses staffing three highly sophisticated ICUs with physiological sensors to identify and prevent exceptionally stressful workflows; and use physiological sensors placed on jet pilots to better train them against fatigue.

 

Sandeep Rath

Kenan Flagler Business School, University of North Carolina at Chapel Hill

Date: Friday, November 17, 2023
Time: 10:00 am - 11:30 am
Location: CSOM 1-122

 

"Collaborative Care for Mental and Physical Health"

Abstract: 
About 27% of patients with diabetes also suffer from depression, and the presence of comorbid depression could increase the cost of care for diabetes by up to 100%. Several randomized clinical trials have demonstrated that physical and mental health are more likely to improve for diabetes patients suffering from depression when regular treatment for depression is provided in a primary care setting (called Collaborative Care). Important operational levers in managing Collaborative Care are the staffing level and allocation of the care manager’s time to enrolled patients based on their requirements. This staffing level and workload allocation influences the revenue, costs, and patient health outcomes. We present a mathematical modeling approach that determines the optimal staffing level and allocation of the care manager’s time and quantifies the costs and benefits of Collaborative Care. In particular, we model Collaborative Care management at the clinic level as an infinite horizon Markov Dynamic Program. The objective is a weighted sum of total patient QALYs and the clinic profits. The model incorporates insurance payment, resource utilization costs, and disease progression of comorbid diabetes and depression. We derive structural properties for the joint optimization of the staffing level and the allocation of care managers’ time to different patient categories. Using these structural properties, we develop a practical and easy-to-implement policy for staffing level and care managers’ time allocation that performs close to the optimal solution. We calibrate the model with data obtained from a large academic medical center and show that our solutions can improve total QALYs and clinic profits when compared to current practices. We also perform sensitivity analysis to different payment models to derive insights relevant to healthcare policy.

Authors: Sandeep Rath, Jayashankar Swaminathan, Charles Coleman

The Supply Chain and Operations Department & Juran Research Center Seminar Series showcases current research in Management and Operations Science, including topics in operations strategy, technology, quality, new product development, and supply chain management. Speakers are drawn from universities around the world.

 

2022 - 2023- Fall

 

Seyed Emadi, Kenan-Flagler School of Business, University of North Carolina           

Date: Friday, November 11, 2022

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

"Diamonds in the Rough: Leveraging Click Data to Spotlight Underrated Products"

Abstract:

Problem Definition. Inspired by a data set from the Chinese retailer JD.com, we study the click and purchase behavior of customers in an online retail setting by employing a structural estimation approach. In particular, by laying out a framework that can disentangle the drivers of customer click versus purchase decisions, we provide insights that can help with product assortment planning.

With the boom in e-commerce, which has been further fueled due to the COVID-19 pandemic, an ever-increasing number of customers are shopping online. Such a rapid growth in online shopping provides retailers with abundance of data to study customer shopping behavior. For example, online retailers can benefit from a better understanding of the customer search (click) and purchase behavior to improve their operational decisions such as assortment planning. From an academic standpoint, even though there is an extensive body of work on consumer search models, there is no prior work (to the best of our knowledge) that separately estimates the products’ attractiveness before and after the click by combining the click and order data.

Methodology/Results. Using a large data set from JD.com, we propose a structural model to estimate the click and purchase behavior of customers according to a dynamic discrete choice model. In particular, we assume that the customer’s utility from each product has an observed and (pre-click) unobserved part (in addition to a random shock). The observed part of the utility is known to the customer prior to the click; however, the unobserved part of the utility can only be learned after the customer clicks on the product. We consider a dynamic program to model the customer’s optimal search strategy. Due to the curse of dimensionality, we propose a novel value function approximation scheme inspired by the Conditional Choice Probability approach. This reduces the estimation to a computationally tractable two-stage process.

By combining the click and order data, our proposed structural framework allows us to disentangle and estimate the observed and unobserved parts of product utilities. Our estimation results show that the value of click for customers can be significant. This is evidenced by the fact that the unobserved utilities of products vary significantly across products. Most importantly, we are able to identify underrated products which we call diamonds in the rough: these are products with low “face” values (i.e., low observed utilities), but high total utilities due to their high (pre-click) unobserved utilities. Thus, even though such products have a low chance of being clicked (due to their low observed utilities), they have a high chance of being purchased, if clicked.

Managerial Implications. Our structural framework provides an online retailer with new tools and insights to better manage the product assortment based on customer click and purchase behavior. In particular, our structural model allows the retailer to disentangle the observed and unobserved parts of product utilities and identify underrated diamond-in-the-rough products. The online retailer can increase the revenue by bringing such products into the spotlight by promoting them on the search page or using tags such as “spotlight product” (similar to “Amazon’s Choice” tags on Amazon.com) to entice customers to click on them. Through simulation studies, we illustrate how our model can improve the assortment decisions by accounting for the unobserved product utilities and significantly increase the revenue (37% on average) compared to an MNL model that only focuses on the observed product utilities.

 

Spring

 

Maxime Cohen, McGill

Date: Friday, January 20, 2023

Time: 10:00 am - 11:30 am

Location: Virtual

"Incentivizing Healthy Food Choices Using Add-on Bundling: A Field Experiment"

Abstract: 

How can retailers incentivize customers to make healthier food choices? Price, convenience, and taste are known to be among the main drivers behind such choices. Unfortunately, healthier food options are often expensive and infrequently promoted. Recent efforts in deploying healthy nudges to incentivize customers toward healthier food choices have been observed. In this paper, we conducted a field experiment with a global convenience store chain to better understand how different add-on bundle promotions influence healthy food choices. We considered three types of add-on bundles: (i) an unhealthy bundle (when customers purchased a coffee, they could add a pastry for $1), (ii) a healthy bundle (offering a healthy snack as an add-on), and (iii) choice bundle (offering either a pastry or a healthy snack). In addition to our field experiment, we conducted an online lab study to strengthen the validity of our results. We found that offering healthy snacks as part of an add-on bundle significantly increased healthy purchases (and decreased unhealthy purchases). Surprisingly, this finding continued to hold for the choice bundle, that is, even when unhealthy snacks were concurrently on promotion. Unfortunately, we did not observe a long-term stickiness effect, meaning that customers returned to their original (unhealthy) purchase patterns once the healthy or choice bundle was discontinued. Finally, we show that offering an add-on choice bundle is also beneficial for retailers, who can earn higher revenue and profit.

 

Shawn Mankad, Cornell University

Date: Friday, February 10, 2023

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

“A Structural Topic Sentiment Model for Text Analysis”

Abstract: 

We consider the common setting where one observes a large number of opinionated text documents and related covariates, such as the text of online reviews along with the date of the review and the author demographic information. In this setting it can be of interest to understand how the covariates determine the text composition as well as the prevalence and sentiment of various discussion themes. Yet, most topic modeling methods are designed to summarize the text for the purpose of exploratory analysis, not to perform this type of formal statistical inference. Further, topic modeling methods generally do not try to estimate the sentiment of discussion along separate topics which can be critical in business applications (e.g., for summarizing service or product quality). We develop a topic model called the Structural Topic Sentiment (STS) model that introduces a new document-level latent sentiment variable for each topic, which modulates the word frequency within a topic. These latent topic sentiment variables are controlled by document-level covariates to allow for experimental control and regression analysis. We also introduce new computational methods to resolve scalability issues that have forced previous models to restrict to a small number of categorical covariates. We benchmark the STS model on three real-world datasets from surveys, blogs, and Yelp restaurant reviews around the coronavirus disease (COVID-19) pandemic. Our model recovers meaningful results including rich insights about how COVID-19 affects online reviews, demonstrating that the STS model can be useful for regression analysis with text data in addition to topic modeling’s traditional use of descriptive analysis.

Jun Li, Stephen M. Ross School of Business, University of Michigan

Date: Friday, April 07, 2023

Time: 10:00 am - 11:30 am

Location: Virtual

“Instrumenting While Experimenting: An Empirical Method for Competitive Pricing at Scale”

Abstract: 

We partner with a leading U.S. e-commerce retailer and develop a competitive pricing method in the context of increasing competition in online retailing. Our method allows retailers to more accurately respond to competitors' price changes at scale. First, we construct a parsimonious demand model that captures the key trade-off in competitive pricing by accounting for two types of customers heterogeneous in their "price-shopping'" behavior. Next, we design and implement a large-scale randomized price experiment on over 10,000 products. Leveraging the experiment as well as the control function approach, we are able to obtain unbiased estimates of the demand model, in particular, price elasticities of both loyal and price-shopping consumers as well as the sales lift when we undercut competitors in price. Lastly, we recommend price responses by solving a constrained optimization problem which uses the estimated demand model as an input. We test this pricing method through another large-scale controlled field experiment on over 10,000 products and demonstrate significant improvements—increasing revenue by over 15% and increasing profit by over 10%. 

 

Santiago Gallino, Wharton School of Business, University of Pennsylvania

Date: Friday, April 28, 2023

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

“Algorithmic Assortment Curation: An Empirical Study of Buy Box in Online Marketplaces”

Abstract:

Most online sales worldwide take place in marketplaces that connect sellers and buyers. The presence of numerous third-party sellers leads to a proliferation of listings for each product, making it difficult for customers to choose between the available options. Online marketplaces adopt algorithmic tools to curate how the different listings for a product are presented to customers. This paper focuses on one such tool, the Buy Box, that algorithmically chooses one option to be presented prominently to customers as a default option. We leveraged the staggered introduction of the Buy Box within a prominent product category in a leading online marketplace to study how the Buy Box impacts marketplace dynamics. Our findings indicate that adopting Buy Box results in a substantial increase in marketplace orders. Implementing Buy Box reduces the frictions customers and sellers face. On the customer side, we find a reduction of search frictions, evidenced by an increase in conversion rates and a higher impact of Buy Box on the mobile channel, which has significantly higher search frictions than desktop channel. On the seller side, the number of sellers offering a product increases following the implementation of Buy Box. Customers benefit from lower prices and higher average quality levels when competition in Buy Boxes is high. After the introduction of the Buy Box, the marketplace also becomes more concentrated. Our paper contributes to the burgeoning literature on the role of algorithms in platforms by examining how algorithmic curation impacts the participants of the marketplace as well as the marketplace dynamics.


 

2021-2022 - Fall

 

Hamsa Bastani, Wharton School, University of Pennsylvania

Date: Friday, October 15th, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Efficient and Targeted COVID-19 Border Testing
via Reinforcement Learning"

Abstract: 

Throughout the COVID-19 pandemic, countries relied on a variety of ad-hoc border control protocols to allow for non-essential travel while safeguarding public health: from quarantining all travelers to restricting entry from select nations based on population-level epidemiological metrics such as cases, deaths or testing positivity rates. Here we report the design and performance of a reinforcement learning system, nicknamed ‘Eva’. In the summer of 2020, Eva was deployed across all Greek borders to limit the influx of asymptomatic travelers infected with SARS-CoV-2, and to inform border policies through real-time estimates of COVID-19 prevalence. In contrast to country-wide protocols, Eva allocated Greece’s limited testing resources based upon incoming travelers’ demographic information and testing results from previous travelers. By comparing Eva’s performance against model good would like to save the power or that's even better put that in the corner thing right away where am Ied counterfactual scenarios, we show that Eva identified 1.85 times as many asymptomatic, infected travelers as random surveillance testing, with up to 2-4 times as many during peak travel, and 1.25-1.45 times as many asymptomatic, infected travelers as testing policies that only utilize epidemiological metrics. We demonstrate that this latter benefit arises, at least partially, because population-level epidemiological metrics had limited predictive value for the actual prevalence of SARS-CoV-2 among asymptomatic travelers and exhibited strong country-specific idiosyncrasies in the summer of 2020. Our results raise serious concerns on the effectiveness of country-agnostic internationally proposed border control policies that are based on population-level epidemiological metrics. Instead, our work represents a successful example of the potential of reinforcement learning and real-time data for safeguarding public health.

 

Spring

 

Bradley Staats, University of North Carolina, Chapel Hill

Date: Friday, February 4, 2022

Time: 9:30 am - 11:00 am

Location: Virtual

"Task Selection and Patient Pick-up: How Familiarity Encourages Physician
Multitasking in the Emergency Department"

Abstract: 

Throughout the COVID-19 pandemic, countries relied on a variety of ad-hoc border control protocols to allow for non-essential travel while safeguarding public health: from quarantining all travellers to restricting entry from select nations based on population-level epidemiological metrics such as cases, deaths or testing positivity rates. Here we report the design and performance of a reinforcement learning system, nicknamed ‘Eva’. In the summer of 2020, Eva was deployed across all Greek borders to limit the influx of asymptomatic travellers infected with SARS-CoV-2, and to inform border policies through real-time estimates of COVID-19 prevalence. In contrast to country-wide protocols, Eva allocated Greece’s limited testing resources based upon incoming travellers’ demographic information and testing results from previous travellers. By comparing Eva’s performance against modelled counterfactual scenarios, we show that Eva identified 1.85 times as many asymptomatic, infected travellers as random surveillance testing, with up to 2-4 times as many during peak travel, and 1.25-1.45 times as many asymptomatic, infected travellers as testing policies that only utilize epidemiological metrics. We demonstrate that this latter benefit arises, at least partially, because population-level epidemiological metrics had limited predictive value for the actual prevalence of SARS-CoV-2 among asymptomatic travellers and exhibited strong country-specific idiosyncrasies in the summer of 2020. Our results raise serious concerns on the effectiveness of country-agnostic internationally proposed border control policies that are based on population-level epidemiological metrics. Instead, our work represents a successful example of the potential of reinforcement learning and real-time data for safeguarding public health.

 

Masha Shunko, University of Washington

Date: Friday, February 11, 2022

Time: 9:30 am - 11:00 am

Location: Virtual

"Evidence of The Unintended Labor Scheduling Implications of The Minimum Wage"

Abstract: 

The effect of the minimum wage is an important yet controversial topic that has received attention for decades. Our study is the first to take an operational lens and empirically study the impact of the minimum wage on firms' scheduling practices. Using a highly granular dataset from a chain of fashion retail stores, we estimate that a $1 increase in the minimum wage, while having a negligible impact on the total labor hours used by the stores, leads to a 27.7% increase in the number of workers scheduled per week, but a 19.4% reduction in weekly hours per worker. For an average store in California, these changes translate into four extra workers and five fewer hours per worker per week. Such scheduling adjustment not only reduces the total wage compensation per worker but also reduces workers' eligibility for benefits. We also show that the minimum wage increase reduces the consistency of weekly and daily schedules for workers. For example, the absolute (relative) deviation in weekly hours worked by each worker increases by up to 32.9% (6.6%) and by up to 9.7% (2.1%) in daily hours, as the minimum wage increases by $1. Our study empirically identifies and highlights a new operational mechanism through which increasing the minimum wage may negatively impact worker welfare. Our further analysis suggests that the combination of the reduced hours, lower eligibility for benefits, and less consistent schedules (that resulted from the minimum wage increase) may substantially hurt worker welfare, even when the overall employment at the stores stay unchanged. By better understanding the intrinsic trade-off of firms' scheduling decisions, policy makers can better design minimum wage policies that will truly benefit workers.

 

Kris (Johnson) Ferreira, Harvard Business School

 

Date: May 6, 2022

Time: 9:30 am -11:00 am

Location: Virtual

"Unlocking Algorithm Potential: Overcoming Naïve Advice Weighting with Feature Transparency"

Abstract:

Although algorithms typically make better forecasts than human decision-makers (HDMs), occasionally HDMs have “private” information which the forecasting algorithm does not have access to - such as social media buzz - that can be used to improve algorithmic forecasts. We first propose a mathematical model that captures how an HDM combines information she directly observes (some of which is private) with an algorithmic prediction to make a final demand forecast. In the model, HDMs take a weighted average between their own forecast and the algorithm’s, where the weights depend only on the aggregate relative historical performances. This leads to HDMs over-adjusting the algorithm’s predictions when it performs well and under-adjusting the algorithm’s predictions when it performs poorly. We validate our model using a lab experiment where 359 participants are tasked with making demand forecasts for 20 products while having access to an algorithm’s recommendations. In a follow-up experiment, we show that providing transparency into the algorithm’s input features can help HDMs use the private information to differentially adjust the algorithm’s forecasts. Our result shows that feature transparency – even when the underlying algorithm is a black box – helps users better incorporate algorithmic recommendations in their decisions.

 

 

2020-2021 - Fall

 

Ariel Stern, Harvard Business School

 

Date: November 13, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

"Product Recalls and New Product Development: Own Firm Distractions
 and Competitor Firm Opportunities”

Abstract:

Product recalls create significant challenges for R&D intensive firms, but simultaneously generate potentially lucrative opportunities for competitors. Using the U.S. medical device industry as our setting, we develop predictions and provide evidence that own firm recalls slow new product development activities, while competitor firm recalls accelerate them. We also examine two firm-level moderators that influence the recall and new product development relationship: product scope and ownership structure. We find that own firm recalls slow new product development for all firm types: a single own firm recall slows new product development up to 43 days, equating to more than $10 million in revenue lost in this high-margin and highly competitive setting. We also find that competitor firm recalls are associated with accelerated development times, but only for broad (vs. narrow) product scope firms and public (vs. private) firms. A one standard deviation increase in competitor firm recalls for these firm types accelerates new product development by more than two weeks. Organizational resources and financial incentives are thus key determinants of whether firms can effectively capitalize on the potential market opportunities created by competitor recalls. In post-hoc analyses, we explore whether future product quality is predicted by post-recall submission times, but find no evidence for this relationship. This result suggests that new product development delays following own firm recalls are more likely driven by organizational distractions than by product quality learning, and that firms react strategically and rationally by speeding new products to market after competitor recalls.

 

Bin Hu, University of Texas, Dallas

 

Date: November 20, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

"Curbing Emissions: Environmental Regulations and
Product Offerings Across Markets" 

Abstract:

In order to curb CO2 emissions, the US Environmental Protection Agency (EPA) implements a minimum federal fuel economy standard for new cars, whereas the California Air Resources Board (CARB) implements a stricter standard in California and other voluntary "CARB states". Automakers have historically met the stricter CARB standard with their offerings. The Trump administration's 2018 announcement to freeze the EPA standard threatened to widen its gap from the CARB standard and cause a split market where automakers offer differentiated car models in CARB and non-CARB states. Inspired by this crisis, we model two regulators---of different levels of environmental awareness---who set efficiency standards in their respective markets, and a firm offering product(s) for these markets. We find that there exist equilibria where the firm offers a unified product for both markets, a different product in each market, or only serves one market, and show that a unified market is preferable to a split market in terms of total emissions. We then propose and show horizontal negotiations (between the regulators) and vertical negotiations (between a regulator and the firm) to be effective strategies to unify a split market and reduce emissions. Finally, we analyze a model variation for local pollutants such as NOx and SO2, as opposed to global pollutants such as CO2, and find that different mitigating strategies may be needed to manage local pollutants.

 

Hessam Bavafa, University of Wisconsin

Date: December 11, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

“The Variance Learning Curve”

Abstract:

The expansive learning curve literature in Operations Management has established how various facets of prior experience improve average performance. In this paper, we explore how increased cumulative experience affects performance variability, or consistency. We use a two-stage estimation method of a heteroskedastic learning curve model to examine the relationship between experience and performance variability among paramedics at the London Ambulance Service. We find that for paramedics with lower experience, an increase in experience of 500 jobs reduces the variance of task completion time by 8.7%, in addition to improving average completion times by 2.7%. Similar to prior results on the average learning curve, we find a diminishing impact of additional experience on the variance learning curve. We provide an evidence base for how to model the learning benefits of cumulative experience on performance in service systems. Our findings imply that the benefits of learning are substantially underestimated if the consistency effect is ignored. Specifically, our estimates indicate that queue lengths (or wait times) might be overestimated by as much as 4% by ignoring the impact of the variance learning curve in service systems. Furthermore, our results suggest that previously established drivers of productivity should be revisited to examine how they affect consistency, in addition to average performance.

 

Spring

Felipe Caro, UCLA, Anderson School of Management

Date: January 22, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Believing in Analytics: Managers' Adherence to
 Price Recommendations from a DSS"

Abstract:

Problem definition: We study the adherence to the recommendations of a decision support system (DSS) for clearance markdowns at Zara, the Spanish fast fashion retailer. Our focus is on behavioral drivers of (i) the decision to deviate from the recommendation, and (ii) the magnitude of the deviation when it occurs. We also study the effect of two related interventions: 1. providing feedback using a revenue metric; and 2. showing a reference point for that metric.  Academic/practical relevance: A major obstacle in the implementation of prescriptive analytics is users' lack of trust in the tool. Understanding the behavioral aspects of managers' usage of these tools is paramount for a successful rollout and deployment.  Methodology: We use data collected by Zara during seven clearance sales campaigns to analyze the operational and behavioral drivers of managers' adherence decisions. Results: Adherence to the DSS's recommendations was higher when such recommendations were aligned with managers' previous coarse heuristic, consistent with algorithm aversion and status quo bias. Adherence was also higher when managers had fewer prices to set, consistent with rational inattention. The magnitude of managers' deviations was larger when inventory levels were higher and sales were slower, and when salvage prices were lower, consistent with the idea that inventory was more salient than revenue, and with loss aversion. Of the two interventions, only the second one was effective in increasing targeted managers' adherence and decreasing their probability of marking down when the DSS recommended leaving a price unchanged. Managerial implications: Our findings provide insights on how to increase voluntary adherence that can be used in any context in which a company wants an analytical tool to be adopted organically by its users. 

This is joint work with Anna Sáez de Tejada Cuenca (IESE Business School).

 

Sridhar Tayur, Carnegie Mellon University, Tepper School of Business

Date: February 5, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

“Quantum Integer Programming (QuIP): An Introduction”

Abstract:  

Many applications in Operations Management, Finance and Cancer Genomics (and others) can be modeled as non-linear (and non-convex) integer programs. Quantum computing, in particular Ising models, provide an alternative way to  tackle these hard problems. In this talk, I will provide an introduction to this new area of Quantum Integer Programming, based on the recent course that I taught at CMU (and was taught at IIT-Madras) last fall, in collaboration with NASA/USRA and Amazon.

 

Nur Sunar, University of North Carolina, Kenan-Flagler Business School

Date: February 12, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Net-Metered Distributed Renewable Energy: A Peril for Utilities?"

 Abstract:

Electricity end-users have been increasingly generating their own electricity via rooftop solar panels. We study the impact of such distributed renewable energy (DRE) on utility profits and social welfare under net metering, which is a widespread policy in the United States. Utilities have been lobbying against net-metered distributed solar based on the common belief that it harms utility profits. We find that when wholesale market dynamics are considered, net-metered DRE may be a positive for utilities. That is, net-metered DRE strictly improves the expected utility profit when the utility’s self-supply is below a threshold and the wholesale electricity price is sufficiently responsive to wholesale demand fluctuations. Our paper distinctively considers both downstream and upstream impacts of net-metered DRE on utilities and analyzes the tradeoff between these impacts. Net-metered DRE can increase utilities’ expenses because of their required buyback from generating customers, and reduces their retail sales revenues. In addition, it can either reduce utilities’ wholesale procurement costs or affect their wholesale market revenues. Our results suggest that utilities might benefit from emerging business strategies that motivate their customers to install solar panels. Our numerical study uses data on the distributed solar in California and the CAISO’s wholesale electricity market and demonstrates that our findings hold under realistic parameters.

 Joint work with Jay Swaminathan

 

Aravind Chandrasekaran, Ohio State University, Fisher College of Business

Date: Friday, February 19, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Does Fresh Food Improve Health?
Expanding the Care Delivery Boundary in Partnership Models of Care"

Abstract:

Problem definition: To address a root cause of poor quality, healthcare institutions are beginning to partner with community-based organizations (CBOs) such as food pantries. These partnership models of care broaden the healthcare delivery infrastructure and, indirectly, couple the distribution of clinical and non-clinical services to increase care continuity. Academic/practical relevance: The healthcare provider identifies the patient’s health-related, social needs and addresses them using a CBO referral. In this paper, we study a partnership model of care called the Fresh Food Farmacy (FFF) program and identify the treatment effect of participation on health quality and cost for those with chronic diseases. The FFF program uses a food insecurity screening tool to individually assess food status and refer patients to nearby food pantries. The partnership model of care advances our understanding of care continuity by expanding the care delivery boundary beyond the healthcare clinic. Methodology: At the patient-clinic visit level of analysis, we operationalize health quality as weight (lbs.) and total cost of care as the summed Medicaid reimbursement rate for primary care procedure codes. We tracked changes in health quality and cost one-year before and one-year after referral for patients with sufficient clinic health data; otherwise, those that met a patient inclusion protocol. We address differences in the amount of clinical health data and observable covariates using a two-step matching procedure and achieve a nearly balanced sample. Econometrically, we use difference-in-differences to identify the treatment effect on health quality and cost. Our total sample includes 1,179 patients with 9,974 clinic visits from 2015 to 2018 from a network of 8 federally-qualified primary care clinics. Results: For patients that comply with the referral, we use a unique identifier to track food pantry access behaviors one-year after referral. We find that participation in the FFF program confers a quality and cost benefit for a subset of patients that accessed the food pantry almost once per month. Among this group, there was a 3% reduction in weight and a 10% reduction in costs. These reductions are clinically significant and suggest that patients have changed underlying health behaviors, a root cause of poor quality. A post-hoc analysis explores the frequency and consistency of food pantry access after referral as a potential explanation for heterogeneous effects. A patient with poor health (at the time of referral) is more likely to use “food as medicine” and benefit more from participating in the FFF program. Managerial implications: Extending the care delivery boundary to include CBOs offers a favorable quality-cost benefit, however, only for certain patients. Healthcare institutions should implement concurrent interventions that target area level barriers to patient participation. 

 

Javad Nasiry, McGill University, Montreal

Date: March 5, 2021

Time: 9:30 am - 11:OO am

Location: Virtual

"Sustainability in the Fast Fashion Industry"

Abstract:

We establish a much-needed link between the fast fashion business model and its environmental consequences. A fast fashion system allows firms to react quickly to changing consumer demand by replenishing inventory (via quick response) and introducing more fashion styles. We study the environmental impact of the fast fashion business model by analyzing its implications for product quality, variety, and inventory decisions. We find that a key driver of low product quality in the fast fashion industry is the firm's incentive to offer variety to hedge against uncertain fashion trends. When variety is endogenous, quality decreases as consumers become more sensitive to fashion or as the cost of introducing new styles decreases. We assess the effectiveness of three environmental initiatives (waste disposal regulations, consumer education, and production tax schemes) in countering the environmental impact of fast fashion. We show that waste disposal policies and production taxes are effective in reducing the firm's leftover inventory--but may have the unintended consequence of lowering product quality and hence worsening the firm's environmental impact.  We also find that education campaigns that increase consumers' sensitivity to quality strictly benefit the environment in the long run.

 

Greys Sosic, University of Southern California, Marshall School of Business

Date: March 12, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Rethinking Salt Supply Chains: Costs and Emissions Analysis
for Co-production of Salt and Freshwater from U.S. Seawater"

Abstract: 

Is it feasible to build desalination plants for the co-production of salt and freshwater from U.S. seawater that could lead to a restructuring of supply chains for salt imports? As water is becoming scarcer worldwide, an increasing number of countries are using desalination plants to generate freshwater. In most such cases, residual concentrates must be disposed off, and the disposal cost is increasing as countries are becoming more environmentally conscious.  Selective salt recovery can help to alleviate this issue, as it reduces the need for concentrate disposal and generates additional revenue. To gain some insights into the costs and benefits of co-production plants, we have collected  data on current desalination practices and salt imports in the U.S., along with the manufacturing costs and energy requirements for co-production plants. We have  used this data to build an optimization model  to determine an optimal number and location of co-production plants in the U.S. and their potential markets for the sale of co-produced salt. In our analysis, we have considered a different total number of co-production facilities, and for each configuration we evaluated the resulting net water cost and carbon emissions impact. Our results indicate that there exists the potential for building several co-production plants in the U.S. that would be both financially competitive with  existing desalination plants and  lead to a reduction in carbon emissions. This information might be of use to both governments and businesses   when they make decisions about the type of desalination facilities  built and the implemented ``polluter pays'' policies.

 

Baris Ata, University of Chicago, Booth School of Business

Date: March 19, 2021

Time: 9:30 am -  11:00 am

Location: Virtual

"Structural Estimation of Kidney Transplant Candidates' Quality of Life Scores:
Improving National Kidney Allocation Policy Under Endogenous Patient Choice and Geographical Structure"

Abstract:

This paper develops a framework for assessing the impact of changes to the deceased-donor kidney allocation policy, taking into account the transplant candidates' (endogenous) organ acceptance behavior. To be specific, it advances a dynamic structural model of the transplant candidates' accept/reject decisions for organ offers. Our formulation models the national list (and its geographic structure) which is important for practical implementations, e.g. for incorporating it in the Kidney Pancreas Simulation Allocation Model (KPSAM). Moreover, it allows various important features of the transplant system such as the degree of tissue matching between the donor and the transplant candidate, changes in the health status of the transplant candidates as they wait on the list, organ quality, geographical sharing and cold-ischemia time of the organs as well as the heterogeneity in transplant candidates' quality of life scores. Using United Network of Organ Sharing (UNOS) data on transplant candidates, donors, organ offers, and follow up results on transplant outcomes, we first estimate the transplant candidates' quality of life scores.  Our estimates are based on patient's revealed preferences and yield similar results on average to what is typically assumed in the medical literature. However, they differ significantly when patient and donor characteristics are considered. We then perform various counterfactual studies for assessing the (unintended) consequences of policy changes. In particular, we find that although the current policy increases the total number of transplants by 2.63% and total life years by 4.45%, it decreases total quality adjusted life years by 1.68%. Moreover, it increases the disparity in probability of getting a transplant for patients of different health scores by 69.3%. These happen due to the current prioritization of healthier patients for kidneys of better quality.  We also show that geographical redistricting of the transplant system, as done for the liver allocation system, does not change the system performance significantly.  However, the brevity matching policy which is a last-come-first served distribution policy based on the health scores of the patients, can further increase the total number of transplants by 1.50%.

 Joint work with: 

John J. Friedewald, Northwestern University Transplant Outcomes Research Center
A. Cem Randa, University of California San Francisco Department of Surgery

 

Ravi Subramanian, Georgia Tech, Scheller College of Business

Date: April 16, 2021

Time: 9:00 am - 11:00 am

Location: Virtual

"An Empirical Investigation of Relationships Between
Locational Demographics and Facility-Level Emissions"

Abstract:

Environmental Justice encompasses the idea of fairness in protecting individuals and communities from environmental and health hazards, regardless of race, color, national origin, or income. Environmental Justice is relevant to the practice of Operations Management in the form of inequities – whether advertent or inadvertent – that result from spatially disparate operational decisions or policies. The broad research question that we aim to address is: How disparate are facility-level emissions and environmental practices across communities with different racial makeups? To empirically address this question, we draw data from the US Census Bureau’s American Community Survey (ACS), and the US EPA’s Toxics Release Inventory (TRI) and Risk-Screening Environmental Indicators Model (RSEI). We employ Coarsened Exact Matching to assess how facilities may differ in their toxicity and exposure-weighted aggregate emissions (outcome) between locations that differ in racial makeup (treatment). Our findings offer evidence for regulatory intervention and opportunities for firms to reframe their broad ESG objectives with local considerations of fairness and equity.

 

Samantha Keppler, University of Michigan

Date: April 23, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

“Crowdfunding the Front Lines: An Empirical Study of
Teacher-Driven School Improvement”

Abstract:

A widespread belief is that the traditional brick-and-mortar K--12 education system in the US is broken. The private sector, specifically education technology (EdTech) companies, have stepped in to try to help. In this paper, we study DonorsChoose, a nonprofit that works to improve the traditional brick-and-mortar system with a teacher crowdfunding platform. Given the constraints of working with the current struggling system, we ask whether DonorsChoose moves the needle on effectiveness and inequality. Combining DonorsChoose data with data on student test scores in Pennsylvania from 2012--2013 to 2017--2018, we find an increase in the number of DonorsChoose projects funded at a school leads to higher student performance, after controlling for selection biases. In high schools, a 10% increase in the number of funded projects leads to a 0.1 to 0.2 percentage point (pp) increase in students scoring basic and above in all tested subjects. A 10% increase in the number of funded projects at an elementary or middle school leads to a 0.06 pp increase in the percentage of students scoring basic and above in language arts and a 0.15 pp increase in science. We find these effects are driven primarily by teacher projects from the lowest income schools, suggesting the platform helps reduce inequality in educational outcomes. Based on a textual analysis of thousands of statements from all funded teachers describing how resources are used, we find two channels of improvement uniquely effective in the lowest income schools. Our study suggests that those in the education sector can harness the wisdom of front-line workers -- teachers -- to improve effectiveness, efficiency, and equity.

 

2019 - 2020 - Spring

Ahmet Colak, Clemson University

Date: March 6, 2020

Time: 9:30 am - 11:00 am

Location: CSOM - 1-132

"Bilateral R&D Productivity and Supply Chain Networks"

Abstract:

We study research and development productivity (RDP) transmission between 4,123 global firms across three supply chain tiers. Collecting 153,090 yearly supply chain dyad partnerships from Bloomberg, we construct a two-sided econometric model of supply chain R&D. In our empirical specification, the dependent variable measures return on R&D, and the independent variables measure supply chain partner and network effects. In our sample data, we find that a 1% R&D productivity improvement of (i) an upstream partner can increase a downstream agent’s R&D productivity by 0.14%, and (ii) a downstream partner can increase an upstream agent’s R&D productivity by 0.28%. Our findings show that having R&D-productive partners plays a significant role in transforming an agent’s R&D into revenues. Similarly, we estimate a network’s average R&D productivity elasticity on an agent as 0.23%. We further find that R&D productivity spreads more within smaller, integrated, domestic, and intra-industry networks. In our two-stage estimation, we address supply chain network endogeneity resulting from entanglement, simultaneity, and partner selection. Our findings provide operational and financial insights for R&D practitioners.

 

2018 - 2019 - Fall

Dennis Cook, University of Minnesota, School of Statistics

Date: Friday, September 28, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 1-132

Title: Partial Least Squares Regression

Abstract: Partial least squares regression, which has been around for about four decades, is a dimension-reduction algorithm for fitting linear regression models without requiring that the sample size be larger than the number of predictors.  It was developed primarily by the Chemometrics community where it is now ingrained as a core method, and it is apparently used throughout the applied sciences.

And yet it seems fair to conclude that PLS regression has not been embraced by some communities, even as a serviceable method that might be useful occasionally. Nor does there seem to be a common understanding as to why this rather enigmatic method should not be used, although bumptious discussions of PLS failings can be found in some applied areas.  Perhaps this is as it should be — perhaps not.

This talk is intended as a relatively informal overview on PLS regression, including historical context, personal encounters, methodology, relationship to envelopes and, near the end, a few recent asymptotic results for high-dimensional regressions.

 

Gopesh Anand, Illinois, Gies College of Business

Date: Friday, October 26, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

 

Tim Kraft, Massachusetts Institute of Technology

Date: Friday, October 5, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

Title: Supply Chain Transparency and Social Responsibility 

Abstract: Consumers increasingly expect companies to ensure that their products are made in a socially responsible manner. However, most companies do not have extensive visibility into their supply chains. According to a recent study, 81% of the 1,700 companies surveyed lacked full visibility into the social responsibility (SR) practices of their suppliers. In this talk, I will first overview our work on how improved supply chain visibility into suppliers’ SR practices can impact companies’ interactions with both consumers and suppliers. I will then discuss a specific study (below) that examines the effect of visibility on consumers’ trust of companies’ SR disclosures.

According to a 2015 Nielsen survey, brand trust tops the list of factors that influence socially responsible purchases. At the same time, supply chain transparency has been identified as one of the most effective ways for companies to improve consumers’ trust of SR disclosures. The current literature focuses on the effect that disclosing information has on consumer trust, generally assuming that companies have complete information (i.e., full supply chain visibility) about the SR practices occurring in their supply chains (e.g., working conditions). In this study, we design an incentivized human-subject experiment to examine whether and how visibility impacts consumers’ trust in companies’ SR communications. Our results show that by investing to improve visibility into the SR practices in its supply chain, a company can increase consumer trust in its SR communications. This increased trust can in turn help the company to increase its sales when a good SR claim is made. This is particularly true among consumers with prosocial orientations; i.e., a willingness to sacrifice their own benefit to help others. 

This talk is based on joint work with Leόn Valdés (University of Pittsburgh) and Karen Zheng (MIT).

 

Ruomeng Cui, Emory University, Goizueta Business School

 

Date: Friday, November 16, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

 

Basak Kalkanci, Georgia Institute of Technology, Scheller College of Business

Date: Friday, November 30, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

Spring

Chris Tang, University of California, Los Angeles 

Date: Friday, May 17, 2019

Time: 10:00-11:30 a.m.

Location: 2-224 CSOM

Title: Environmental Violations in China: Descriptive, Predictive, and Prescriptive Analysis

Abstract:  Air and water pollution in China has crated public concern.  In this talk, we discuss our empirical findings about the short-term and long-term implications of Chinese manufacturers who violated environmental regulations.   Also, we present a model that can enable Chinese government to use publicly available data to predict which firm is more likely to violate regulations in the future.

*This talk is based on two research papers: 1. Chris KY Lo, Christopher S Tang, Yi Zhou, Andy CL Yeung, Di Fan, 2018. Environmental Incidents and the Market Value of Firms: An Empirical Investigation in the Chinese Context, M&SOM; and 2. Chris KY Lo, Christopher S Tang, Yi Zhou, 2018. Environmental Violations in China: Evaluating Their Long-term Impact and Predicting Future Violations, Working Paper, UCLA.  


2017 - 2018 - Fall

Susan Lu, Purdue, Krannert School of Management

Date: Friday, October 6, 2017

Time: 10:00-11:30 a.m.

Location: CSOM 1-132

Title: Do Mandatory Overtime Law Improve Quality? Staffing Decisions and Operational Flexibility of Nursing Homes

Abstract: TBD

 

Soo-Haeng Cho, Carnegie Mellon University, Tepper School of Business

Date: Friday, October 27, 2017

Time: 10:00-11:30 a.m.

Location: CSOM 1-135

Title: The Theory of Crowdsourcing Contests

Abstract: 

     This talk will present papers that concern the theory of innovation tournaments (also called crowdsourcing contests or open innovation). In an innovation tournament, an organizer solicits innovative ideas from a number of independent agents. Agents exert efforts to develop their solutions, but their outcomes are unknown due to technical uncertainty and/or subjective evaluation criteria. We call an agent whose ex-post solution contributes to the organizer's utility a “contributor.”

     The first paper, entitled “Optimal Award Scheme in Innovation Tournaments,” examines optimal award scheme/rule in innovation tournaments. While extant literature either assumes a winner-take-all scheme a priori or shows its optimality under specific distributions for uncertainty, this paper derives necessary and sufficient conditions under which the winner-take-all scheme is optimal. These conditions are violated when agents perceive it very likely that only few agents receive high evaluation or when a tournament does not require substantial increase in agents' marginal cost of effort to develop high-quality solutions. Yet, the winner-take-all scheme is optimal in many practical situations, especially when agents have symmetric beliefs about their evaluation. In this case, the organizer should offer a larger winner prize when he is interested in obtaining a higher number of good solutions, but interestingly the organizer need not necessarily raise the winner prize when anticipating more participants to a tournament.

     The second paper, entitled “Innovation Tournaments with Multiple Contributors,” analyzes a general model of uncertainty and utility functions with multiple contributors, and shows that these factors play a crucial role in decision-making of agents and the organizer. Specifically, contrary to existing theories, increased competition to a tournament can have a positive impact on agents' incentives to exert effort when agents expect good outcomes with high likelihood, and a free-entry open tournament should be encouraged only when the problem is highly uncertain or the organizer seeks diverse solutions from many contributors. Our results are consistent with recent empirical evidence, hence helping close a gap in the extant literature between theory and practice.

 

Spring

John Birge, The University of Chicago, Booth School of Business

Date: Friday, March 30

Time: 10:00-11:30 a.m.

Location: CSOM 2-224

Title: Dynamic Learning in Strategic Pricing Games

Abstract:

In monopoly pricing situations, firms should optimally vary prices to learn demand. The variation must be sufficiently high to ensure complete learning.  In competitive situations, however, varying prices provides information to competitors and may reduce the value of learning. Such situations may arise in the pricing of new products such as pharmaceuticals. This talk will discuss how this effect can be strong enough to stop learning so that firms optimally reduce any variation in prices and choose not to learn demand. The result can be that the selling firms achieve a collaborative outcome instead of a competitive equilibrium. The result has implications for policies that restrict price changes or require disclosures. 

 

Karen Zheng, MIT, Sloan

Date: Friday, April 27

Time: 10:00-11:30 a.m.

Location: CSOM 2-213

Title: Economically Motivated Adulteration in Farming Supply Chains: Data and Models

Abstract: 

Food adulteration is a serious threat to public health. Many incidents of food adulteration are motivated by economic gains, defined as economically motivated adulteration (EMA). In this talk, I will present recent works that examine drivers of EMA risks in farming supply chains. We first present a multi-industry empirical study that demonstrates supply chain dispersion and local governance being two important risk drivers. To do so, we collect farming supply chain data, food sampling data, and socioeconomic data across five different industries (aquatic products, eggs, honey, pork, and poultry) in China. We define supply chain dispersion as the degree to which farming outputs are sourced from a dispersed network of farms -- and develop a method to quantify dispersion in farming supply chains based on field data. We also develop multi-faceted measures to objectively quantify the strength of city-level governance in China based on factual (as opposed to perception) data. Our results show that products made in a more dispersed supply chain and in regions with weaker governance are associated with higher EMA risks. Inspired by the empirical findings, we develop a set of supply network models to structurally analyze farms’ strategic adulteration behavior and the resulting EMA risks in farming supply chains. We examine both preemptive EMA, where farms engage in adulteration to reduce the likelihood of producing low-quality units, and reactive EMA, where farms adulterate low-quality units to increase the perceived quality of their outputs. We fully characterize the farms’ equilibrium adulteration behavior in both types of EMA and examine how quality uncertainty, supply chain dispersion, traceability, and testing sensitivity (in detecting adulteration) jointly impact EMA risks. We validate our models’ predictions with numerical analyses calibrated by field data. We conclude by offering tangible insights regarding how business entities and regulators can more proactively prevent and address EMA risks in farming supply chains.

The Supply Chain and Operations Department & Juran Research Center Seminar Series showcases current research in Management and Operations Science, including topics in operations strategy, technology, quality, new product development, and supply chain management. Speakers are drawn from universities around the world.


2023-2024 Spring

Jonathan Eugene Helm

Kelley School of Business, Indiana University

Date: Friday, February 9, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

Abstract:
The opioid crisis has ravaged the United States, taking 69,000 lives in 2020, with prescription opioids accounting for 98% of opioid abuse. Though this epidemic is often considered a White public health crisis nationally, overdose deaths have doubled among people of color from 2017-2019. Significant public criticism has gone to several individual firms involved as research has demonstrated that the opioid crisis was driven by excessive supply from the pharmaceutical industry. We find evidence that the scope of the blame goes beyond individual actors to include the very structure of supply chains, where complex supply chains may have exacerbated the crisis by dispensing significantly more opioids. We posit that supply chain complexity allowed mass quantities of opioids to escape detection by the Drug Enforcement Administration (DEA). Further, we find new evidence showing the greater impact of complexity on dispensing in non-White communities, which underscores their exclusion from the public discourse and governmental response surrounding the crisis and suggests possible racial bias in the DEA’s regulatory policies. Our analysis was made possible by the hotly contested 2019 release of the DEA’s ARCOS database that logs every shipment in the supply chain of opioids in the United States from 2006-2014. Using a fixed effects model, we find that a one-unit increase across three dimensions of supply chain complexity is associated with a 16% increase in opioid dispensing. This effect is intensified in non-White communities, where a 10% increase in the non-White population is associated with a 3.39% (1.33%) increase in the effects of complexity for high (average) complexity supply chains. To verify that excess dispensing of high-complexity pharmacies supplied non-medical/recreational demand, we exploit the reformulation of OxyContin designed to prevent recreational use as an exogenous shock to the market. In a novel approach, we use the fact that different pharmacies received their first shipment of reformulated OxyContin at different times and use a difference-in-differences model to estimate the heterogeneous effect of the shock on dispensing. As the abuse-deterrent OxyContin stifled demand, high complexity pharmacies experienced a 15.31% greater reduction in dispensing compared to lower-complexity pharmacies, suggesting the excess dispensing was satisfying non-medical/recreational demand.

 

Michelle Shell

The Tuck School of Business, Dartmouth

Date: Friday, April 5. 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132

 

Saed Alizamir

Darden School of Business, University of Virginia

Date: Friday, April 19, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

 

Telesilla Kotsi

Fisher College of Business, Ohio State

Date: Friday, May 10, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132

 

Ashish Kabra

Robert H. Smith School of Business, University of Maryland

Date: Friday, May 24, 2024
Time: 10:00 am - 11:30 am
Location: CSOM 1-132


2023-2024 Fall

 

Suvrat Dhanorkar

Smeal College of Business, Penn State

Date: Friday, September 22, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"Does Legalizing Marijuana Increase Toxic Waste?"

Evidence from Manufacturing Facilities in the United States"

Abstract:
Problem Definition: Marijuana is increasingly gaining legal acceptance in the United States. From 1996 to 2016, 28 states and the District of Columbia legalized marijuana for medical use. At the same time, studies find that when individuals use marijuana, they can experience several negative consequences such as reduced attention, hampered memory, and poor task completion—which implies that marijuana can degrade the output of human resources. But human resources are a critical input into manufacturing and are vital for ensuring operational effectiveness, which suggests that increased access to marijuana could adversely affect manufacturing operations. However, hardly any work has explored the impact of legalizing marijuana on manufacturing activities. In this study, we seek to bridge the gap by examining the effect of legalizing marijuana on the toxic chemical releases of manufacturing facilities.

Methodology/Results: We leverage a state-level quasi-experimental setting that evolves from the staggered enactment of marijuana legislation by different states in the U.S. We augment this setup with data from the Environmental Protection Agency’s Toxics Release Inventory program. Our dataset spans 1987-2016 and includes details on the toxic chemical releases of 45,720 manufacturing facilities in the U.S. We find that medical marijuana legislation (MML) adversely affects the toxic releases of facilities in the state—the average releases increased by 5.22% after MML. Further examination reveals that facilities undertake fewer managerial and technical modifications to their operational processes, which clarifies the mechanisms that affect the toxic releases. Finally, we find that recreational marijuana legislation (RML), which increases access to marijuana for the general population, leads to further increases of toxic releases—the effect goes beyond the impact of MML.

Managerial Implications: Our study provides important insights for managers and policy makers by casting light on the detrimental impact of legalizing marijuana on manufacturing operations.

 

Priyank Arora

Darla Moore School of Business, University of South Carolina

Date: Friday, October 6, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"Farm Equipment Sharing in Emerging Economies"

Abstract:
In emerging economies, a growing number of farm equipment sharing platforms have emerged to connect smallholder farmers with tractor owners who are willing to fulfill farmers’ requests for mechanization services. Due to the small farm sizes and the low digital literacy in rural areas of emerging economies, these platforms often rely on the so-called “booking agents” to collect demand from individual farmers and submit the aggregated demand on the platform. This is in contrast to other conventional sharing platforms where no booking agents are present and the service providers directly fulfill service requests from individual customers. In this paper, we explicitly capture the role of booking agents on a farm equipment sharing platform and study how the platform should choose the price and wage rates to appropriately incentivize all entities on the platform (including farmers, tractor owners, and booking agents). Although the farm equipment sharing platforms are popularly referred to as “Uber for Tractors,” our analysis offers insights about when and why these platforms should not rely on conventional wisdom derived from other conventional sharing settings (such as ride-sharing), due to the presence of booking agents. Further, inspired by the increasing efforts in practice from governments or donor agencies to enhance the supply and demand sides of these platforms (such as by increasing the number of tractors on the platform or by making booking agents’ demand aggregation more efficient), our analysis also sheds light on how such efforts affect the platforms’ optimal decisions and the equilibrium outcomes. Finally, we present a calibrated numerical study using data from Hello Tractor—an award-winning farm equipment sharing platform—to provide an illustration of how the generated insights can map to practice.

*Joint work with: Olufunke Adebola (Deloitte; previously, Hello Tractor) and Can Zhang (Fuqua School of Business, Duke University)

 

Meng Li

Bauer School of Business, University of Houston

Date: Friday, October 27, 2023
Time: 10:00 am - 11:30 am
Location: CSOM L-122

 

"The Value of AI in OM: Evidence from Field Experiments"

Abstract:
I will discuss two papers. In the first paper “AI and Procurement,” we study how buyers’ use of artificial intelligence (AI) affects suppliers’ price quoting strategies. Specifically, we study the impact of automation—that is, the buyer uses a chatbot to automatically inquire about prices instead of asking in person—and the impact of smartness—that is, the buyer signals the use of a smart AI algorithm in selecting the supplier. We collaborate with a trading company to run a field experiment on an online platform in which we compare suppliers’ wholesale price quotes across female, male, and chatbot buyer types under AI and no recommendation conditions. We find that, when not equipped with a smart control, there is price discrimination against chatbot buyers who receive a higher wholesale price quote than human buyers. In fact, without smartness, automation alone receives the highest quoted wholesale price. However, signaling the use of a smart recommendation system can effectively reduce suppliers’ price quote for chatbot buyers. We also show that AI delivers the most value when buyers adopt automation and smartness simultaneously in procurement.

In the second paper “Physician Adoption of AI Assistant,” we study AI assistants---software agents that can perform tasks or services for individuals---which are among the most promising AI applications. However, little is known about the adoption of AI assistants by service providers (i.e., physicians) in a real-world healthcare setting. In this paper, we investigate the impact of AI smartness (i.e., whether the AI assistant is empowered by machine learning intelligence) and the impact of AI transparency (i.e., whether physicians are informed of the AI assistant). We collaborate with a leading healthcare platform to run a field experiment in which we compare physicians’ adoption behavior, i.e., adoption rate and adoption timing, of smart and automated AI assistants under transparent and non-transparent conditions. We find that smartness can increase the adoption rate and shorten the adoption timing, while transparency can only shorten the adoption timing. Moreover, the impact of AI transparency on the adoption rate is contingent on the smartness level of the AI assistant: the transparency increases the adoption rate only when the AI assistant is not equipped with smart algorithms and fails to do so when the AI assistant is smart. Our study can guide platforms in designing their AI strategies. Platforms should improve the smartness of AI assistant. If such an improvement is too costly, the platform should transparentize the AI assistant, especially when it is not smart.

 

Ken Moon

Wharton School, University Of Pennsylvania 

Date: Friday, November 3, 2023
Time: 10:00 am - 11:30 am
Location: CSOM 1-142

 

"Bringing Data Science to the Management of Workforces"

Abstract:
The talk will cover several, real-world collaborations relating to the operational management of workforces.  The main part of the talk will focus on a research project with the Apple Worker Exit Study using extensive data on staffing, productivity, and pay from within a consumer electronics supply chain producing tens of billions in USD revenue quarterly.  We study how firms should manage the problem of worker turnover, including its surprising impact on low-skilled workforces and the implications for production, wage, and inventory decisions.  Despite the lack of skills, we find that worker turnover impedes coordination between assembly line coworkers by weakening knowledge sharing and relationships.  We structurally estimate a dynamic equilibrium model of workers’ endogenous turnover decisions and the firm’s dynamic production and staffing decisions, and we apply reinforcement learning to evaluate managerial alternatives.  A less turnover-prone, hence more productive, workforce reduces the firm’s variable production costs by 4.5%, or an estimated $928 million for the studied product. Such benefits justify paying higher efficiency wages even to less skilled workforces; furthermore, interestingly, rational inventory management policies incentivize self-interested firms to reduce rather than tolerate turnover.   We also cover more recent research that develops learning algorithms to address the problem of worker stress and burnout for highly skilled workforces (ICU nurses and fighter jet pilots).  In particular, we equip the nurses staffing three highly sophisticated ICUs with physiological sensors to identify and prevent exceptionally stressful workflows; and use physiological sensors placed on jet pilots to better train them against fatigue.

 

Sandeep Rath

Kenan Flagler Business School, University of North Carolina at Chapel Hill

Date: Friday, November 17, 2023
Time: 10:00 am - 11:30 am
Location: CSOM 1-122

 

"Collaborative Care for Mental and Physical Health"

Abstract: 
About 27% of patients with diabetes also suffer from depression, and the presence of comorbid depression could increase the cost of care for diabetes by up to 100%. Several randomized clinical trials have demonstrated that physical and mental health are more likely to improve for diabetes patients suffering from depression when regular treatment for depression is provided in a primary care setting (called Collaborative Care). Important operational levers in managing Collaborative Care are the staffing level and allocation of the care manager’s time to enrolled patients based on their requirements. This staffing level and workload allocation influences the revenue, costs, and patient health outcomes. We present a mathematical modeling approach that determines the optimal staffing level and allocation of the care manager’s time and quantifies the costs and benefits of Collaborative Care. In particular, we model Collaborative Care management at the clinic level as an infinite horizon Markov Dynamic Program. The objective is a weighted sum of total patient QALYs and the clinic profits. The model incorporates insurance payment, resource utilization costs, and disease progression of comorbid diabetes and depression. We derive structural properties for the joint optimization of the staffing level and the allocation of care managers’ time to different patient categories. Using these structural properties, we develop a practical and easy-to-implement policy for staffing level and care managers’ time allocation that performs close to the optimal solution. We calibrate the model with data obtained from a large academic medical center and show that our solutions can improve total QALYs and clinic profits when compared to current practices. We also perform sensitivity analysis to different payment models to derive insights relevant to healthcare policy.

Authors: Sandeep Rath, Jayashankar Swaminathan, Charles Coleman

The Supply Chain and Operations Department & Juran Research Center Seminar Series showcases current research in Management and Operations Science, including topics in operations strategy, technology, quality, new product development, and supply chain management. Speakers are drawn from universities around the world.

 

2022 - 2023- Fall

 

Seyed Emadi, Kenan-Flagler School of Business, University of North Carolina           

Date: Friday, November 11, 2022

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

"Diamonds in the Rough: Leveraging Click Data to Spotlight Underrated Products"

Abstract:

Problem Definition. Inspired by a data set from the Chinese retailer JD.com, we study the click and purchase behavior of customers in an online retail setting by employing a structural estimation approach. In particular, by laying out a framework that can disentangle the drivers of customer click versus purchase decisions, we provide insights that can help with product assortment planning.

With the boom in e-commerce, which has been further fueled due to the COVID-19 pandemic, an ever-increasing number of customers are shopping online. Such a rapid growth in online shopping provides retailers with abundance of data to study customer shopping behavior. For example, online retailers can benefit from a better understanding of the customer search (click) and purchase behavior to improve their operational decisions such as assortment planning. From an academic standpoint, even though there is an extensive body of work on consumer search models, there is no prior work (to the best of our knowledge) that separately estimates the products’ attractiveness before and after the click by combining the click and order data.

Methodology/Results. Using a large data set from JD.com, we propose a structural model to estimate the click and purchase behavior of customers according to a dynamic discrete choice model. In particular, we assume that the customer’s utility from each product has an observed and (pre-click) unobserved part (in addition to a random shock). The observed part of the utility is known to the customer prior to the click; however, the unobserved part of the utility can only be learned after the customer clicks on the product. We consider a dynamic program to model the customer’s optimal search strategy. Due to the curse of dimensionality, we propose a novel value function approximation scheme inspired by the Conditional Choice Probability approach. This reduces the estimation to a computationally tractable two-stage process.

By combining the click and order data, our proposed structural framework allows us to disentangle and estimate the observed and unobserved parts of product utilities. Our estimation results show that the value of click for customers can be significant. This is evidenced by the fact that the unobserved utilities of products vary significantly across products. Most importantly, we are able to identify underrated products which we call diamonds in the rough: these are products with low “face” values (i.e., low observed utilities), but high total utilities due to their high (pre-click) unobserved utilities. Thus, even though such products have a low chance of being clicked (due to their low observed utilities), they have a high chance of being purchased, if clicked.

Managerial Implications. Our structural framework provides an online retailer with new tools and insights to better manage the product assortment based on customer click and purchase behavior. In particular, our structural model allows the retailer to disentangle the observed and unobserved parts of product utilities and identify underrated diamond-in-the-rough products. The online retailer can increase the revenue by bringing such products into the spotlight by promoting them on the search page or using tags such as “spotlight product” (similar to “Amazon’s Choice” tags on Amazon.com) to entice customers to click on them. Through simulation studies, we illustrate how our model can improve the assortment decisions by accounting for the unobserved product utilities and significantly increase the revenue (37% on average) compared to an MNL model that only focuses on the observed product utilities.

 

Spring

 

Maxime Cohen, McGill

Date: Friday, January 20, 2023

Time: 10:00 am - 11:30 am

Location: Virtual

"Incentivizing Healthy Food Choices Using Add-on Bundling: A Field Experiment"

Abstract: 

How can retailers incentivize customers to make healthier food choices? Price, convenience, and taste are known to be among the main drivers behind such choices. Unfortunately, healthier food options are often expensive and infrequently promoted. Recent efforts in deploying healthy nudges to incentivize customers toward healthier food choices have been observed. In this paper, we conducted a field experiment with a global convenience store chain to better understand how different add-on bundle promotions influence healthy food choices. We considered three types of add-on bundles: (i) an unhealthy bundle (when customers purchased a coffee, they could add a pastry for $1), (ii) a healthy bundle (offering a healthy snack as an add-on), and (iii) choice bundle (offering either a pastry or a healthy snack). In addition to our field experiment, we conducted an online lab study to strengthen the validity of our results. We found that offering healthy snacks as part of an add-on bundle significantly increased healthy purchases (and decreased unhealthy purchases). Surprisingly, this finding continued to hold for the choice bundle, that is, even when unhealthy snacks were concurrently on promotion. Unfortunately, we did not observe a long-term stickiness effect, meaning that customers returned to their original (unhealthy) purchase patterns once the healthy or choice bundle was discontinued. Finally, we show that offering an add-on choice bundle is also beneficial for retailers, who can earn higher revenue and profit.

 

Shawn Mankad, Cornell University

Date: Friday, February 10, 2023

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

“A Structural Topic Sentiment Model for Text Analysis”

Abstract: 

We consider the common setting where one observes a large number of opinionated text documents and related covariates, such as the text of online reviews along with the date of the review and the author demographic information. In this setting it can be of interest to understand how the covariates determine the text composition as well as the prevalence and sentiment of various discussion themes. Yet, most topic modeling methods are designed to summarize the text for the purpose of exploratory analysis, not to perform this type of formal statistical inference. Further, topic modeling methods generally do not try to estimate the sentiment of discussion along separate topics which can be critical in business applications (e.g., for summarizing service or product quality). We develop a topic model called the Structural Topic Sentiment (STS) model that introduces a new document-level latent sentiment variable for each topic, which modulates the word frequency within a topic. These latent topic sentiment variables are controlled by document-level covariates to allow for experimental control and regression analysis. We also introduce new computational methods to resolve scalability issues that have forced previous models to restrict to a small number of categorical covariates. We benchmark the STS model on three real-world datasets from surveys, blogs, and Yelp restaurant reviews around the coronavirus disease (COVID-19) pandemic. Our model recovers meaningful results including rich insights about how COVID-19 affects online reviews, demonstrating that the STS model can be useful for regression analysis with text data in addition to topic modeling’s traditional use of descriptive analysis.

Jun Li, Stephen M. Ross School of Business, University of Michigan

Date: Friday, April 07, 2023

Time: 10:00 am - 11:30 am

Location: Virtual

“Instrumenting While Experimenting: An Empirical Method for Competitive Pricing at Scale”

Abstract: 

We partner with a leading U.S. e-commerce retailer and develop a competitive pricing method in the context of increasing competition in online retailing. Our method allows retailers to more accurately respond to competitors' price changes at scale. First, we construct a parsimonious demand model that captures the key trade-off in competitive pricing by accounting for two types of customers heterogeneous in their "price-shopping'" behavior. Next, we design and implement a large-scale randomized price experiment on over 10,000 products. Leveraging the experiment as well as the control function approach, we are able to obtain unbiased estimates of the demand model, in particular, price elasticities of both loyal and price-shopping consumers as well as the sales lift when we undercut competitors in price. Lastly, we recommend price responses by solving a constrained optimization problem which uses the estimated demand model as an input. We test this pricing method through another large-scale controlled field experiment on over 10,000 products and demonstrate significant improvements—increasing revenue by over 15% and increasing profit by over 10%. 

 

Santiago Gallino, Wharton School of Business, University of Pennsylvania

Date: Friday, April 28, 2023

Time: 10:00 am - 11:30 am

Location: CSOM 2-213

“Algorithmic Assortment Curation: An Empirical Study of Buy Box in Online Marketplaces”

Abstract:

Most online sales worldwide take place in marketplaces that connect sellers and buyers. The presence of numerous third-party sellers leads to a proliferation of listings for each product, making it difficult for customers to choose between the available options. Online marketplaces adopt algorithmic tools to curate how the different listings for a product are presented to customers. This paper focuses on one such tool, the Buy Box, that algorithmically chooses one option to be presented prominently to customers as a default option. We leveraged the staggered introduction of the Buy Box within a prominent product category in a leading online marketplace to study how the Buy Box impacts marketplace dynamics. Our findings indicate that adopting Buy Box results in a substantial increase in marketplace orders. Implementing Buy Box reduces the frictions customers and sellers face. On the customer side, we find a reduction of search frictions, evidenced by an increase in conversion rates and a higher impact of Buy Box on the mobile channel, which has significantly higher search frictions than desktop channel. On the seller side, the number of sellers offering a product increases following the implementation of Buy Box. Customers benefit from lower prices and higher average quality levels when competition in Buy Boxes is high. After the introduction of the Buy Box, the marketplace also becomes more concentrated. Our paper contributes to the burgeoning literature on the role of algorithms in platforms by examining how algorithmic curation impacts the participants of the marketplace as well as the marketplace dynamics.


 

2021-2022 - Fall

 

Hamsa Bastani, Wharton School, University of Pennsylvania

Date: Friday, October 15th, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Efficient and Targeted COVID-19 Border Testing
via Reinforcement Learning"

Abstract: 

Throughout the COVID-19 pandemic, countries relied on a variety of ad-hoc border control protocols to allow for non-essential travel while safeguarding public health: from quarantining all travelers to restricting entry from select nations based on population-level epidemiological metrics such as cases, deaths or testing positivity rates. Here we report the design and performance of a reinforcement learning system, nicknamed ‘Eva’. In the summer of 2020, Eva was deployed across all Greek borders to limit the influx of asymptomatic travelers infected with SARS-CoV-2, and to inform border policies through real-time estimates of COVID-19 prevalence. In contrast to country-wide protocols, Eva allocated Greece’s limited testing resources based upon incoming travelers’ demographic information and testing results from previous travelers. By comparing Eva’s performance against model good would like to save the power or that's even better put that in the corner thing right away where am Ied counterfactual scenarios, we show that Eva identified 1.85 times as many asymptomatic, infected travelers as random surveillance testing, with up to 2-4 times as many during peak travel, and 1.25-1.45 times as many asymptomatic, infected travelers as testing policies that only utilize epidemiological metrics. We demonstrate that this latter benefit arises, at least partially, because population-level epidemiological metrics had limited predictive value for the actual prevalence of SARS-CoV-2 among asymptomatic travelers and exhibited strong country-specific idiosyncrasies in the summer of 2020. Our results raise serious concerns on the effectiveness of country-agnostic internationally proposed border control policies that are based on population-level epidemiological metrics. Instead, our work represents a successful example of the potential of reinforcement learning and real-time data for safeguarding public health.

 

Spring

 

Bradley Staats, University of North Carolina, Chapel Hill

Date: Friday, February 4, 2022

Time: 9:30 am - 11:00 am

Location: Virtual

"Task Selection and Patient Pick-up: How Familiarity Encourages Physician
Multitasking in the Emergency Department"

Abstract: 

Throughout the COVID-19 pandemic, countries relied on a variety of ad-hoc border control protocols to allow for non-essential travel while safeguarding public health: from quarantining all travellers to restricting entry from select nations based on population-level epidemiological metrics such as cases, deaths or testing positivity rates. Here we report the design and performance of a reinforcement learning system, nicknamed ‘Eva’. In the summer of 2020, Eva was deployed across all Greek borders to limit the influx of asymptomatic travellers infected with SARS-CoV-2, and to inform border policies through real-time estimates of COVID-19 prevalence. In contrast to country-wide protocols, Eva allocated Greece’s limited testing resources based upon incoming travellers’ demographic information and testing results from previous travellers. By comparing Eva’s performance against modelled counterfactual scenarios, we show that Eva identified 1.85 times as many asymptomatic, infected travellers as random surveillance testing, with up to 2-4 times as many during peak travel, and 1.25-1.45 times as many asymptomatic, infected travellers as testing policies that only utilize epidemiological metrics. We demonstrate that this latter benefit arises, at least partially, because population-level epidemiological metrics had limited predictive value for the actual prevalence of SARS-CoV-2 among asymptomatic travellers and exhibited strong country-specific idiosyncrasies in the summer of 2020. Our results raise serious concerns on the effectiveness of country-agnostic internationally proposed border control policies that are based on population-level epidemiological metrics. Instead, our work represents a successful example of the potential of reinforcement learning and real-time data for safeguarding public health.

 

Masha Shunko, University of Washington

Date: Friday, February 11, 2022

Time: 9:30 am - 11:00 am

Location: Virtual

"Evidence of The Unintended Labor Scheduling Implications of The Minimum Wage"

Abstract: 

The effect of the minimum wage is an important yet controversial topic that has received attention for decades. Our study is the first to take an operational lens and empirically study the impact of the minimum wage on firms' scheduling practices. Using a highly granular dataset from a chain of fashion retail stores, we estimate that a $1 increase in the minimum wage, while having a negligible impact on the total labor hours used by the stores, leads to a 27.7% increase in the number of workers scheduled per week, but a 19.4% reduction in weekly hours per worker. For an average store in California, these changes translate into four extra workers and five fewer hours per worker per week. Such scheduling adjustment not only reduces the total wage compensation per worker but also reduces workers' eligibility for benefits. We also show that the minimum wage increase reduces the consistency of weekly and daily schedules for workers. For example, the absolute (relative) deviation in weekly hours worked by each worker increases by up to 32.9% (6.6%) and by up to 9.7% (2.1%) in daily hours, as the minimum wage increases by $1. Our study empirically identifies and highlights a new operational mechanism through which increasing the minimum wage may negatively impact worker welfare. Our further analysis suggests that the combination of the reduced hours, lower eligibility for benefits, and less consistent schedules (that resulted from the minimum wage increase) may substantially hurt worker welfare, even when the overall employment at the stores stay unchanged. By better understanding the intrinsic trade-off of firms' scheduling decisions, policy makers can better design minimum wage policies that will truly benefit workers.

 

Kris (Johnson) Ferreira, Harvard Business School

 

Date: May 6, 2022

Time: 9:30 am -11:00 am

Location: Virtual

"Unlocking Algorithm Potential: Overcoming Naïve Advice Weighting with Feature Transparency"

Abstract:

Although algorithms typically make better forecasts than human decision-makers (HDMs), occasionally HDMs have “private” information which the forecasting algorithm does not have access to - such as social media buzz - that can be used to improve algorithmic forecasts. We first propose a mathematical model that captures how an HDM combines information she directly observes (some of which is private) with an algorithmic prediction to make a final demand forecast. In the model, HDMs take a weighted average between their own forecast and the algorithm’s, where the weights depend only on the aggregate relative historical performances. This leads to HDMs over-adjusting the algorithm’s predictions when it performs well and under-adjusting the algorithm’s predictions when it performs poorly. We validate our model using a lab experiment where 359 participants are tasked with making demand forecasts for 20 products while having access to an algorithm’s recommendations. In a follow-up experiment, we show that providing transparency into the algorithm’s input features can help HDMs use the private information to differentially adjust the algorithm’s forecasts. Our result shows that feature transparency – even when the underlying algorithm is a black box – helps users better incorporate algorithmic recommendations in their decisions.

 

 

2020-2021 - Fall

 

Ariel Stern, Harvard Business School

 

Date: November 13, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

"Product Recalls and New Product Development: Own Firm Distractions
 and Competitor Firm Opportunities”

Abstract:

Product recalls create significant challenges for R&D intensive firms, but simultaneously generate potentially lucrative opportunities for competitors. Using the U.S. medical device industry as our setting, we develop predictions and provide evidence that own firm recalls slow new product development activities, while competitor firm recalls accelerate them. We also examine two firm-level moderators that influence the recall and new product development relationship: product scope and ownership structure. We find that own firm recalls slow new product development for all firm types: a single own firm recall slows new product development up to 43 days, equating to more than $10 million in revenue lost in this high-margin and highly competitive setting. We also find that competitor firm recalls are associated with accelerated development times, but only for broad (vs. narrow) product scope firms and public (vs. private) firms. A one standard deviation increase in competitor firm recalls for these firm types accelerates new product development by more than two weeks. Organizational resources and financial incentives are thus key determinants of whether firms can effectively capitalize on the potential market opportunities created by competitor recalls. In post-hoc analyses, we explore whether future product quality is predicted by post-recall submission times, but find no evidence for this relationship. This result suggests that new product development delays following own firm recalls are more likely driven by organizational distractions than by product quality learning, and that firms react strategically and rationally by speeding new products to market after competitor recalls.

 

Bin Hu, University of Texas, Dallas

 

Date: November 20, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

"Curbing Emissions: Environmental Regulations and
Product Offerings Across Markets" 

Abstract:

In order to curb CO2 emissions, the US Environmental Protection Agency (EPA) implements a minimum federal fuel economy standard for new cars, whereas the California Air Resources Board (CARB) implements a stricter standard in California and other voluntary "CARB states". Automakers have historically met the stricter CARB standard with their offerings. The Trump administration's 2018 announcement to freeze the EPA standard threatened to widen its gap from the CARB standard and cause a split market where automakers offer differentiated car models in CARB and non-CARB states. Inspired by this crisis, we model two regulators---of different levels of environmental awareness---who set efficiency standards in their respective markets, and a firm offering product(s) for these markets. We find that there exist equilibria where the firm offers a unified product for both markets, a different product in each market, or only serves one market, and show that a unified market is preferable to a split market in terms of total emissions. We then propose and show horizontal negotiations (between the regulators) and vertical negotiations (between a regulator and the firm) to be effective strategies to unify a split market and reduce emissions. Finally, we analyze a model variation for local pollutants such as NOx and SO2, as opposed to global pollutants such as CO2, and find that different mitigating strategies may be needed to manage local pollutants.

 

Hessam Bavafa, University of Wisconsin

Date: December 11, 2020

Time: 9:30 am - 11:00 am

Location: Virtual

“The Variance Learning Curve”

Abstract:

The expansive learning curve literature in Operations Management has established how various facets of prior experience improve average performance. In this paper, we explore how increased cumulative experience affects performance variability, or consistency. We use a two-stage estimation method of a heteroskedastic learning curve model to examine the relationship between experience and performance variability among paramedics at the London Ambulance Service. We find that for paramedics with lower experience, an increase in experience of 500 jobs reduces the variance of task completion time by 8.7%, in addition to improving average completion times by 2.7%. Similar to prior results on the average learning curve, we find a diminishing impact of additional experience on the variance learning curve. We provide an evidence base for how to model the learning benefits of cumulative experience on performance in service systems. Our findings imply that the benefits of learning are substantially underestimated if the consistency effect is ignored. Specifically, our estimates indicate that queue lengths (or wait times) might be overestimated by as much as 4% by ignoring the impact of the variance learning curve in service systems. Furthermore, our results suggest that previously established drivers of productivity should be revisited to examine how they affect consistency, in addition to average performance.

 

Spring

Felipe Caro, UCLA, Anderson School of Management

Date: January 22, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Believing in Analytics: Managers' Adherence to
 Price Recommendations from a DSS"

Abstract:

Problem definition: We study the adherence to the recommendations of a decision support system (DSS) for clearance markdowns at Zara, the Spanish fast fashion retailer. Our focus is on behavioral drivers of (i) the decision to deviate from the recommendation, and (ii) the magnitude of the deviation when it occurs. We also study the effect of two related interventions: 1. providing feedback using a revenue metric; and 2. showing a reference point for that metric.  Academic/practical relevance: A major obstacle in the implementation of prescriptive analytics is users' lack of trust in the tool. Understanding the behavioral aspects of managers' usage of these tools is paramount for a successful rollout and deployment.  Methodology: We use data collected by Zara during seven clearance sales campaigns to analyze the operational and behavioral drivers of managers' adherence decisions. Results: Adherence to the DSS's recommendations was higher when such recommendations were aligned with managers' previous coarse heuristic, consistent with algorithm aversion and status quo bias. Adherence was also higher when managers had fewer prices to set, consistent with rational inattention. The magnitude of managers' deviations was larger when inventory levels were higher and sales were slower, and when salvage prices were lower, consistent with the idea that inventory was more salient than revenue, and with loss aversion. Of the two interventions, only the second one was effective in increasing targeted managers' adherence and decreasing their probability of marking down when the DSS recommended leaving a price unchanged. Managerial implications: Our findings provide insights on how to increase voluntary adherence that can be used in any context in which a company wants an analytical tool to be adopted organically by its users. 

This is joint work with Anna Sáez de Tejada Cuenca (IESE Business School).

 

Sridhar Tayur, Carnegie Mellon University, Tepper School of Business

Date: February 5, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

“Quantum Integer Programming (QuIP): An Introduction”

Abstract:  

Many applications in Operations Management, Finance and Cancer Genomics (and others) can be modeled as non-linear (and non-convex) integer programs. Quantum computing, in particular Ising models, provide an alternative way to  tackle these hard problems. In this talk, I will provide an introduction to this new area of Quantum Integer Programming, based on the recent course that I taught at CMU (and was taught at IIT-Madras) last fall, in collaboration with NASA/USRA and Amazon.

 

Nur Sunar, University of North Carolina, Kenan-Flagler Business School

Date: February 12, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Net-Metered Distributed Renewable Energy: A Peril for Utilities?"

 Abstract:

Electricity end-users have been increasingly generating their own electricity via rooftop solar panels. We study the impact of such distributed renewable energy (DRE) on utility profits and social welfare under net metering, which is a widespread policy in the United States. Utilities have been lobbying against net-metered distributed solar based on the common belief that it harms utility profits. We find that when wholesale market dynamics are considered, net-metered DRE may be a positive for utilities. That is, net-metered DRE strictly improves the expected utility profit when the utility’s self-supply is below a threshold and the wholesale electricity price is sufficiently responsive to wholesale demand fluctuations. Our paper distinctively considers both downstream and upstream impacts of net-metered DRE on utilities and analyzes the tradeoff between these impacts. Net-metered DRE can increase utilities’ expenses because of their required buyback from generating customers, and reduces their retail sales revenues. In addition, it can either reduce utilities’ wholesale procurement costs or affect their wholesale market revenues. Our results suggest that utilities might benefit from emerging business strategies that motivate their customers to install solar panels. Our numerical study uses data on the distributed solar in California and the CAISO’s wholesale electricity market and demonstrates that our findings hold under realistic parameters.

 Joint work with Jay Swaminathan

 

Aravind Chandrasekaran, Ohio State University, Fisher College of Business

Date: Friday, February 19, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Does Fresh Food Improve Health?
Expanding the Care Delivery Boundary in Partnership Models of Care"

Abstract:

Problem definition: To address a root cause of poor quality, healthcare institutions are beginning to partner with community-based organizations (CBOs) such as food pantries. These partnership models of care broaden the healthcare delivery infrastructure and, indirectly, couple the distribution of clinical and non-clinical services to increase care continuity. Academic/practical relevance: The healthcare provider identifies the patient’s health-related, social needs and addresses them using a CBO referral. In this paper, we study a partnership model of care called the Fresh Food Farmacy (FFF) program and identify the treatment effect of participation on health quality and cost for those with chronic diseases. The FFF program uses a food insecurity screening tool to individually assess food status and refer patients to nearby food pantries. The partnership model of care advances our understanding of care continuity by expanding the care delivery boundary beyond the healthcare clinic. Methodology: At the patient-clinic visit level of analysis, we operationalize health quality as weight (lbs.) and total cost of care as the summed Medicaid reimbursement rate for primary care procedure codes. We tracked changes in health quality and cost one-year before and one-year after referral for patients with sufficient clinic health data; otherwise, those that met a patient inclusion protocol. We address differences in the amount of clinical health data and observable covariates using a two-step matching procedure and achieve a nearly balanced sample. Econometrically, we use difference-in-differences to identify the treatment effect on health quality and cost. Our total sample includes 1,179 patients with 9,974 clinic visits from 2015 to 2018 from a network of 8 federally-qualified primary care clinics. Results: For patients that comply with the referral, we use a unique identifier to track food pantry access behaviors one-year after referral. We find that participation in the FFF program confers a quality and cost benefit for a subset of patients that accessed the food pantry almost once per month. Among this group, there was a 3% reduction in weight and a 10% reduction in costs. These reductions are clinically significant and suggest that patients have changed underlying health behaviors, a root cause of poor quality. A post-hoc analysis explores the frequency and consistency of food pantry access after referral as a potential explanation for heterogeneous effects. A patient with poor health (at the time of referral) is more likely to use “food as medicine” and benefit more from participating in the FFF program. Managerial implications: Extending the care delivery boundary to include CBOs offers a favorable quality-cost benefit, however, only for certain patients. Healthcare institutions should implement concurrent interventions that target area level barriers to patient participation. 

 

Javad Nasiry, McGill University, Montreal

Date: March 5, 2021

Time: 9:30 am - 11:OO am

Location: Virtual

"Sustainability in the Fast Fashion Industry"

Abstract:

We establish a much-needed link between the fast fashion business model and its environmental consequences. A fast fashion system allows firms to react quickly to changing consumer demand by replenishing inventory (via quick response) and introducing more fashion styles. We study the environmental impact of the fast fashion business model by analyzing its implications for product quality, variety, and inventory decisions. We find that a key driver of low product quality in the fast fashion industry is the firm's incentive to offer variety to hedge against uncertain fashion trends. When variety is endogenous, quality decreases as consumers become more sensitive to fashion or as the cost of introducing new styles decreases. We assess the effectiveness of three environmental initiatives (waste disposal regulations, consumer education, and production tax schemes) in countering the environmental impact of fast fashion. We show that waste disposal policies and production taxes are effective in reducing the firm's leftover inventory--but may have the unintended consequence of lowering product quality and hence worsening the firm's environmental impact.  We also find that education campaigns that increase consumers' sensitivity to quality strictly benefit the environment in the long run.

 

Greys Sosic, University of Southern California, Marshall School of Business

Date: March 12, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

"Rethinking Salt Supply Chains: Costs and Emissions Analysis
for Co-production of Salt and Freshwater from U.S. Seawater"

Abstract: 

Is it feasible to build desalination plants for the co-production of salt and freshwater from U.S. seawater that could lead to a restructuring of supply chains for salt imports? As water is becoming scarcer worldwide, an increasing number of countries are using desalination plants to generate freshwater. In most such cases, residual concentrates must be disposed off, and the disposal cost is increasing as countries are becoming more environmentally conscious.  Selective salt recovery can help to alleviate this issue, as it reduces the need for concentrate disposal and generates additional revenue. To gain some insights into the costs and benefits of co-production plants, we have collected  data on current desalination practices and salt imports in the U.S., along with the manufacturing costs and energy requirements for co-production plants. We have  used this data to build an optimization model  to determine an optimal number and location of co-production plants in the U.S. and their potential markets for the sale of co-produced salt. In our analysis, we have considered a different total number of co-production facilities, and for each configuration we evaluated the resulting net water cost and carbon emissions impact. Our results indicate that there exists the potential for building several co-production plants in the U.S. that would be both financially competitive with  existing desalination plants and  lead to a reduction in carbon emissions. This information might be of use to both governments and businesses   when they make decisions about the type of desalination facilities  built and the implemented ``polluter pays'' policies.

 

Baris Ata, University of Chicago, Booth School of Business

Date: March 19, 2021

Time: 9:30 am -  11:00 am

Location: Virtual

"Structural Estimation of Kidney Transplant Candidates' Quality of Life Scores:
Improving National Kidney Allocation Policy Under Endogenous Patient Choice and Geographical Structure"

Abstract:

This paper develops a framework for assessing the impact of changes to the deceased-donor kidney allocation policy, taking into account the transplant candidates' (endogenous) organ acceptance behavior. To be specific, it advances a dynamic structural model of the transplant candidates' accept/reject decisions for organ offers. Our formulation models the national list (and its geographic structure) which is important for practical implementations, e.g. for incorporating it in the Kidney Pancreas Simulation Allocation Model (KPSAM). Moreover, it allows various important features of the transplant system such as the degree of tissue matching between the donor and the transplant candidate, changes in the health status of the transplant candidates as they wait on the list, organ quality, geographical sharing and cold-ischemia time of the organs as well as the heterogeneity in transplant candidates' quality of life scores. Using United Network of Organ Sharing (UNOS) data on transplant candidates, donors, organ offers, and follow up results on transplant outcomes, we first estimate the transplant candidates' quality of life scores.  Our estimates are based on patient's revealed preferences and yield similar results on average to what is typically assumed in the medical literature. However, they differ significantly when patient and donor characteristics are considered. We then perform various counterfactual studies for assessing the (unintended) consequences of policy changes. In particular, we find that although the current policy increases the total number of transplants by 2.63% and total life years by 4.45%, it decreases total quality adjusted life years by 1.68%. Moreover, it increases the disparity in probability of getting a transplant for patients of different health scores by 69.3%. These happen due to the current prioritization of healthier patients for kidneys of better quality.  We also show that geographical redistricting of the transplant system, as done for the liver allocation system, does not change the system performance significantly.  However, the brevity matching policy which is a last-come-first served distribution policy based on the health scores of the patients, can further increase the total number of transplants by 1.50%.

 Joint work with: 

John J. Friedewald, Northwestern University Transplant Outcomes Research Center
A. Cem Randa, University of California San Francisco Department of Surgery

 

Ravi Subramanian, Georgia Tech, Scheller College of Business

Date: April 16, 2021

Time: 9:00 am - 11:00 am

Location: Virtual

"An Empirical Investigation of Relationships Between
Locational Demographics and Facility-Level Emissions"

Abstract:

Environmental Justice encompasses the idea of fairness in protecting individuals and communities from environmental and health hazards, regardless of race, color, national origin, or income. Environmental Justice is relevant to the practice of Operations Management in the form of inequities – whether advertent or inadvertent – that result from spatially disparate operational decisions or policies. The broad research question that we aim to address is: How disparate are facility-level emissions and environmental practices across communities with different racial makeups? To empirically address this question, we draw data from the US Census Bureau’s American Community Survey (ACS), and the US EPA’s Toxics Release Inventory (TRI) and Risk-Screening Environmental Indicators Model (RSEI). We employ Coarsened Exact Matching to assess how facilities may differ in their toxicity and exposure-weighted aggregate emissions (outcome) between locations that differ in racial makeup (treatment). Our findings offer evidence for regulatory intervention and opportunities for firms to reframe their broad ESG objectives with local considerations of fairness and equity.

 

Samantha Keppler, University of Michigan

Date: April 23, 2021

Time: 9:30 am - 11:00 am

Location: Virtual

“Crowdfunding the Front Lines: An Empirical Study of
Teacher-Driven School Improvement”

Abstract:

A widespread belief is that the traditional brick-and-mortar K--12 education system in the US is broken. The private sector, specifically education technology (EdTech) companies, have stepped in to try to help. In this paper, we study DonorsChoose, a nonprofit that works to improve the traditional brick-and-mortar system with a teacher crowdfunding platform. Given the constraints of working with the current struggling system, we ask whether DonorsChoose moves the needle on effectiveness and inequality. Combining DonorsChoose data with data on student test scores in Pennsylvania from 2012--2013 to 2017--2018, we find an increase in the number of DonorsChoose projects funded at a school leads to higher student performance, after controlling for selection biases. In high schools, a 10% increase in the number of funded projects leads to a 0.1 to 0.2 percentage point (pp) increase in students scoring basic and above in all tested subjects. A 10% increase in the number of funded projects at an elementary or middle school leads to a 0.06 pp increase in the percentage of students scoring basic and above in language arts and a 0.15 pp increase in science. We find these effects are driven primarily by teacher projects from the lowest income schools, suggesting the platform helps reduce inequality in educational outcomes. Based on a textual analysis of thousands of statements from all funded teachers describing how resources are used, we find two channels of improvement uniquely effective in the lowest income schools. Our study suggests that those in the education sector can harness the wisdom of front-line workers -- teachers -- to improve effectiveness, efficiency, and equity.

 

2019 - 2020 - Spring

Ahmet Colak, Clemson University

Date: March 6, 2020

Time: 9:30 am - 11:00 am

Location: CSOM - 1-132

"Bilateral R&D Productivity and Supply Chain Networks"

Abstract:

We study research and development productivity (RDP) transmission between 4,123 global firms across three supply chain tiers. Collecting 153,090 yearly supply chain dyad partnerships from Bloomberg, we construct a two-sided econometric model of supply chain R&D. In our empirical specification, the dependent variable measures return on R&D, and the independent variables measure supply chain partner and network effects. In our sample data, we find that a 1% R&D productivity improvement of (i) an upstream partner can increase a downstream agent’s R&D productivity by 0.14%, and (ii) a downstream partner can increase an upstream agent’s R&D productivity by 0.28%. Our findings show that having R&D-productive partners plays a significant role in transforming an agent’s R&D into revenues. Similarly, we estimate a network’s average R&D productivity elasticity on an agent as 0.23%. We further find that R&D productivity spreads more within smaller, integrated, domestic, and intra-industry networks. In our two-stage estimation, we address supply chain network endogeneity resulting from entanglement, simultaneity, and partner selection. Our findings provide operational and financial insights for R&D practitioners.

 

2018 - 2019 - Fall

Dennis Cook, University of Minnesota, School of Statistics

Date: Friday, September 28, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 1-132

Title: Partial Least Squares Regression

Abstract: Partial least squares regression, which has been around for about four decades, is a dimension-reduction algorithm for fitting linear regression models without requiring that the sample size be larger than the number of predictors.  It was developed primarily by the Chemometrics community where it is now ingrained as a core method, and it is apparently used throughout the applied sciences.

And yet it seems fair to conclude that PLS regression has not been embraced by some communities, even as a serviceable method that might be useful occasionally. Nor does there seem to be a common understanding as to why this rather enigmatic method should not be used, although bumptious discussions of PLS failings can be found in some applied areas.  Perhaps this is as it should be — perhaps not.

This talk is intended as a relatively informal overview on PLS regression, including historical context, personal encounters, methodology, relationship to envelopes and, near the end, a few recent asymptotic results for high-dimensional regressions.

 

Gopesh Anand, Illinois, Gies College of Business

Date: Friday, October 26, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

 

Tim Kraft, Massachusetts Institute of Technology

Date: Friday, October 5, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

Title: Supply Chain Transparency and Social Responsibility 

Abstract: Consumers increasingly expect companies to ensure that their products are made in a socially responsible manner. However, most companies do not have extensive visibility into their supply chains. According to a recent study, 81% of the 1,700 companies surveyed lacked full visibility into the social responsibility (SR) practices of their suppliers. In this talk, I will first overview our work on how improved supply chain visibility into suppliers’ SR practices can impact companies’ interactions with both consumers and suppliers. I will then discuss a specific study (below) that examines the effect of visibility on consumers’ trust of companies’ SR disclosures.

According to a 2015 Nielsen survey, brand trust tops the list of factors that influence socially responsible purchases. At the same time, supply chain transparency has been identified as one of the most effective ways for companies to improve consumers’ trust of SR disclosures. The current literature focuses on the effect that disclosing information has on consumer trust, generally assuming that companies have complete information (i.e., full supply chain visibility) about the SR practices occurring in their supply chains (e.g., working conditions). In this study, we design an incentivized human-subject experiment to examine whether and how visibility impacts consumers’ trust in companies’ SR communications. Our results show that by investing to improve visibility into the SR practices in its supply chain, a company can increase consumer trust in its SR communications. This increased trust can in turn help the company to increase its sales when a good SR claim is made. This is particularly true among consumers with prosocial orientations; i.e., a willingness to sacrifice their own benefit to help others. 

This talk is based on joint work with Leόn Valdés (University of Pittsburgh) and Karen Zheng (MIT).

 

Ruomeng Cui, Emory University, Goizueta Business School

 

Date: Friday, November 16, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

 

Basak Kalkanci, Georgia Institute of Technology, Scheller College of Business

Date: Friday, November 30, 2018

Time: 10:00-11:30 a.m.

Location: CSOM 2-233

Spring

Chris Tang, University of California, Los Angeles 

Date: Friday, May 17, 2019

Time: 10:00-11:30 a.m.

Location: 2-224 CSOM

Title: Environmental Violations in China: Descriptive, Predictive, and Prescriptive Analysis

Abstract:  Air and water pollution in China has crated public concern.  In this talk, we discuss our empirical findings about the short-term and long-term implications of Chinese manufacturers who violated environmental regulations.   Also, we present a model that can enable Chinese government to use publicly available data to predict which firm is more likely to violate regulations in the future.

*This talk is based on two research papers: 1. Chris KY Lo, Christopher S Tang, Yi Zhou, Andy CL Yeung, Di Fan, 2018. Environmental Incidents and the Market Value of Firms: An Empirical Investigation in the Chinese Context, M&SOM; and 2. Chris KY Lo, Christopher S Tang, Yi Zhou, 2018. Environmental Violations in China: Evaluating Their Long-term Impact and Predicting Future Violations, Working Paper, UCLA.  


2017 - 2018 - Fall

Susan Lu, Purdue, Krannert School of Management

Date: Friday, October 6, 2017

Time: 10:00-11:30 a.m.

Location: CSOM 1-132

Title: Do Mandatory Overtime Law Improve Quality? Staffing Decisions and Operational Flexibility of Nursing Homes

Abstract: TBD

 

Soo-Haeng Cho, Carnegie Mellon University, Tepper School of Business

Date: Friday, October 27, 2017

Time: 10:00-11:30 a.m.

Location: CSOM 1-135

Title: The Theory of Crowdsourcing Contests

Abstract: 

     This talk will present papers that concern the theory of innovation tournaments (also called crowdsourcing contests or open innovation). In an innovation tournament, an organizer solicits innovative ideas from a number of independent agents. Agents exert efforts to develop their solutions, but their outcomes are unknown due to technical uncertainty and/or subjective evaluation criteria. We call an agent whose ex-post solution contributes to the organizer's utility a “contributor.”

     The first paper, entitled “Optimal Award Scheme in Innovation Tournaments,” examines optimal award scheme/rule in innovation tournaments. While extant literature either assumes a winner-take-all scheme a priori or shows its optimality under specific distributions for uncertainty, this paper derives necessary and sufficient conditions under which the winner-take-all scheme is optimal. These conditions are violated when agents perceive it very likely that only few agents receive high evaluation or when a tournament does not require substantial increase in agents' marginal cost of effort to develop high-quality solutions. Yet, the winner-take-all scheme is optimal in many practical situations, especially when agents have symmetric beliefs about their evaluation. In this case, the organizer should offer a larger winner prize when he is interested in obtaining a higher number of good solutions, but interestingly the organizer need not necessarily raise the winner prize when anticipating more participants to a tournament.

     The second paper, entitled “Innovation Tournaments with Multiple Contributors,” analyzes a general model of uncertainty and utility functions with multiple contributors, and shows that these factors play a crucial role in decision-making of agents and the organizer. Specifically, contrary to existing theories, increased competition to a tournament can have a positive impact on agents' incentives to exert effort when agents expect good outcomes with high likelihood, and a free-entry open tournament should be encouraged only when the problem is highly uncertain or the organizer seeks diverse solutions from many contributors. Our results are consistent with recent empirical evidence, hence helping close a gap in the extant literature between theory and practice.

 

Spring

John Birge, The University of Chicago, Booth School of Business

Date: Friday, March 30

Time: 10:00-11:30 a.m.

Location: CSOM 2-224

Title: Dynamic Learning in Strategic Pricing Games

Abstract:

In monopoly pricing situations, firms should optimally vary prices to learn demand. The variation must be sufficiently high to ensure complete learning.  In competitive situations, however, varying prices provides information to competitors and may reduce the value of learning. Such situations may arise in the pricing of new products such as pharmaceuticals. This talk will discuss how this effect can be strong enough to stop learning so that firms optimally reduce any variation in prices and choose not to learn demand. The result can be that the selling firms achieve a collaborative outcome instead of a competitive equilibrium. The result has implications for policies that restrict price changes or require disclosures. 

 

Karen Zheng, MIT, Sloan

Date: Friday, April 27

Time: 10:00-11:30 a.m.

Location: CSOM 2-213

Title: Economically Motivated Adulteration in Farming Supply Chains: Data and Models

Abstract: 

Food adulteration is a serious threat to public health. Many incidents of food adulteration are motivated by economic gains, defined as economically motivated adulteration (EMA). In this talk, I will present recent works that examine drivers of EMA risks in farming supply chains. We first present a multi-industry empirical study that demonstrates supply chain dispersion and local governance being two important risk drivers. To do so, we collect farming supply chain data, food sampling data, and socioeconomic data across five different industries (aquatic products, eggs, honey, pork, and poultry) in China. We define supply chain dispersion as the degree to which farming outputs are sourced from a dispersed network of farms -- and develop a method to quantify dispersion in farming supply chains based on field data. We also develop multi-faceted measures to objectively quantify the strength of city-level governance in China based on factual (as opposed to perception) data. Our results show that products made in a more dispersed supply chain and in regions with weaker governance are associated with higher EMA risks. Inspired by the empirical findings, we develop a set of supply network models to structurally analyze farms’ strategic adulteration behavior and the resulting EMA risks in farming supply chains. We examine both preemptive EMA, where farms engage in adulteration to reduce the likelihood of producing low-quality units, and reactive EMA, where farms adulterate low-quality units to increase the perceived quality of their outputs. We fully characterize the farms’ equilibrium adulteration behavior in both types of EMA and examine how quality uncertainty, supply chain dispersion, traceability, and testing sensitivity (in detecting adulteration) jointly impact EMA risks. We validate our models’ predictions with numerical analyses calibrated by field data. We conclude by offering tangible insights regarding how business entities and regulators can more proactively prevent and address EMA risks in farming supply chains.

Supply Chain & Operations Department