SCAN-BASED TRADING: OPPORTUNITIES FOR ENHANCING SUPPLY-CHAIN EFFICIENCY

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: AFRI COMPETITIVE GRANT
• Reporting Frequency: Annual
• Accession No.: 1011923
• Grant No.: 2017-67023-26236
• Cumulative Award Amt.: $496,407.00
• Proposal No.: 2016-09900
• Project Start Date: Jun 1, 2017
• Project End Date: May 31, 2019
• Grant Year: 2017
• Program Code: [A1641]- Agriculture Economics and Rural Communities: Markets and Trade

Recipient Organization
ARIZONA STATE UNIVERSITY
660 S MILL AVE STE 312
TEMPE,AZ 85281-3670

Performing Department
Supply Chain Management

Non Technical Summary
A resurgence of mergers among food retailers and advances in supply-chain technology have raised concerns regarding retail-buyer market power. Scan-based trading (SBT) is a new type of supply contract in which the supplier retains title until the product is checked-out by the customer. Any loss of product between the delivery and checkout is the responsibility of the supplier, not the retailer. If managed properly, SBT provides many advantages for both suppliers (farmers) and retailers, but the potential for abuse by retailers is clear. Our objective is to gain a better understanding of whether SBT reflects an abuse of market power, or an opportunity for greater supply-chain efficiency. To that end, we derive optimal SBT contracts in a principalagent context that align the incentives of supply chain members. Using field-experiment data from a major bakery-products manufacturer, we empirically evaluate the impact SBT has on shrink, wholesale prices, and service levels. We also estimate a structural model of vertical retail supply relationships under SBT in order to test for differences in bargaining power between SBT and traditional supply contracts. An economic bargaining experiment will be conducted to determine the effect of exogenously varying market power on the shrink-share and pricing terms of SBT contracts. The estimated structural model will be simulated to determine the impact of SBT on welfare outcomes for producers, retailers, and consumers, under various shrink and pricing terms. Our findings will produce new insights into how SBT contracts work, or can work to support the objectives of various stakeholders.

Animal Health Component

100%

Research Effort Categories

Basic

(N/A)

Applied

100%

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
604	6220	3010	100%

Knowledge Area
604 - Marketing and Distribution Practices;

Subject Of Investigation
6220 - Marketing, processing, and supply firms other than cooperatives;

Field Of Science
3010 - Economics;

Keywords

buying power

food waste

market structure

principal-agent

structural model

supply-chain analysis

Goals / Objectives
Our primary objective is to gain a better understanding of whether emerging supply contracts, such as SBT, between food producers and retailers reflect asymmetric vertical market power, or are merely examples of new transactional technology, and a search for efficiency. In order to achieve this primary objective, we will also pursue a number of supporting objectives. To achieve the supporting objectives, we will:1. (Stage 1 of the proposed research) develop a conceptual framework in which we model SBT contracts in the framework of a principal-agent transaction, and derive optimal SBT contracts that align the incentives of suppliers and retailers, and maximize the potential benefits of SBT contracts;2. (Stage 2) complete an empirical analysis of the field-experiment data obtained from a cooperating bakery supplier and Nielsen, including a detailed summary of the data in order to reveal any stylized facts regarding the apparent relationship between the use of SBT and inventory shrink, SBT and wholesale prices, or SBT and the frequency of stock-outs, and estimates of a di¤erence-in-di¤erence empirical model in which we formally test the relationship between SBT contracts and shrink-share implied by the principal-agent model;3. (Stage 3) design and estimate a structural model of vertical retail supply relationships under SBT in which shrink-share and wholesale prices are the key control variables. In this model, we again use the bakery / Nielsen .eld experiment data to identify the relationship between contract form and the relative bargaining power of retailers and the supplier;4. (Stage 4) develop and implement an economic bargaining experiment between buying and selling subjects in order to determine the e¤ect of exogenously varying bargaining power on the shrink-share and pricing terms of SBT contracts;5. (Stage 5) simulate the structural model estimated in Stage 3 in order to determine the impact of SBT contracts, under various shrink and pricing terms, on welfare outcomes for producer / suppliers, retailers, consumers, and society more generally;6. (Stage 6) draw implications of our .ndings for how SBT contracts can be used to advantage by both retailers and suppliers, including the potential to minimize food waste throughout the supply system for our focal category.Achieving each of these objectives will produce a number of substantially new insights into how SBT contracts work, or can work to support the objectives of both food retailers, suppliers, and the agricultural producers who sell to them.

Project Methods
Scan Based Trading (SBT) is an inventory management practice in which the supplier retains title until the product is scanned through checkout by the customer. The supplier, therefore, is responsible for maintaining inventory levels and replenishment schedules, while the retailers' employees manage the shelf-space itself. Importantly, any "shrink" or loss of product between the delivery and checkout is the responsibility of the supplier, not the retailer. In a perfect world, shrink should be a small amount of the total supplied to the store, but in reality it is not. One the greatest problem that SBT creates for suppliers and consumers is "shrink". The fact that suppliers become liable for retailers' behavior at the retail store makes this SBT phenomenon a classic principal-agent problem, an observation that has not been fully explored in the literature. Further, retailers may sell many different brands, including their own private label. Therefore, the retailer's objective is fundamentally misaligned with that of the supplier, and may not manage the supplier's product in the best possible way, perhaps increasing the probability that product will be allowed to perish on the shelf, or be pushed to the side to allow faster-moving, fresher items to sell. In the proposed research, we plan to design a theoretical model of optimal contract design, and test its implications using data from a unique field experiment provided by a major bakery-goods supplier.In our theoretical model, suppliers design contracts by varying the proportion of shrinkage borne by the retailer, or "shrink share," in order to align their objectives. Our theoretical model will recognize that any SBT contract must be both incentive compatible, meaning that it takes into account the utility-maximization incentives of both parties to the contract, and the participation constraint that must be met in order to ensure that both sign the contract. We use this model to derive core hypotheses that highlight the nature of the potential problem with SBT, namely that when we endow retailers with a high level of bargaining power, the shrink share will be near zero, and equilibrium shrink will be higher than under a traditional DSD system.We test hypotheses regarding inventory loss that emerge from our theoretical model of retailer-supplier contracting using a unique data set from one of the largest bakery suppliers in the US, combined with Nielsen ScanTrak retail sales data. The manufacturer maintains a comprehensive dataset that covers supply transactions for their products sold through a number of major retailers. In our case, we have data on both the price and quantity of each item sold, the cost and quantity supplied to the retailer, and inventory levels. We test the effect of any principal-agent problems created by SBT by extracting a natural experiment from the bakery data. That is, we impute the level of shrink on a weekly basis from the bakery data, and estimate a reduced-form model of weekly shrink as a function of whether the retailer is under a SBT or DSD contract, and a number of other control variables. This natural experiment allows us to identify the causal effect of imposing SBT on the level of shrink by comparing the amounts shipped between comparable retailers that do and do not use SBT. By econometrically estimating the SBT-treatment effect, we can control for temporal, brand, and retailer variation that may otherwise confound the clear identification of the SBT effect. From this empirical model, however, we will be able to test for any causal relationships between SBT and shrink, but not the underlying mechanism that may explain why a retailer may agree to an SBT contract.The emergence of SBT raises questions of how bilateral market power affects vertical relationships in the food supply network more generally. When an SBT contract emerges out of a bilateral supply relationship with terms that are particularly unfavorable to the supplier, there is circumstantial evidence that a power asymmetry exists. However, contracts in reality are more complicated than the stylized version we describe in our simple principal-agent model, and prices are likely to be a key element. Therefore, we conduct a second empirical exercise in which we use the bakery data to test for the relationship between SBT contracting the degree of bargaining power exercised by retailers or wholesalers in the vertical channel. Because our combined datasets include both retail and wholesale prices for a set of focal items, it is ideally suited to estimating a structural model of vertical supplier-retailer contracting in which a Nash bargaining power parameter defines the relative share of the total (retail less production cost) margin earned by each party. In this model, we use the SBT / non-SBT experimental data to parameterize the Nash bargaining power parameter into regimes of SBT and non-SBT contracting. In this way, we estimate how vertical bargaining power varies with the type of contract used. In an extension to this structural model, we will also allow shrink share to be a second endogenous variable that conditions bargaining over the equilibrium wholesale price. Data at the level of detail available from our bakery supplier will allow us to identify bargaining power consistently, and accurately, without the missing data problem that commonly bedevils the identification of bargaining power. Our results from this model will show whether the use of SBT contracts is associated with the supplier's exercise of power, or retailers' lack thereof.Any empirical exercise with secondary data, however, is subject to a criticism that there are confounding factors that are not perfectly controlled, or that identification cannot be assured. Therefore, we also design and implement an incentive-compatible economic bargaining experiment in which we are able to control for environmental conditions that may draw our focus away from the relationships we are interested in, namely the exercise of bargaining power and the joint outcomes of retailer shrink-share and wholesale prices. Bargaining experiments are commonly used to study situations in which the market is subject to bilateral variation in bargaining power, or at least some expectation of asymmetry in bargaining power. Our experiment is unique in that buyers and sellers negotiate over contracts with 2 attributes .shrink share and wholesale prices .so by varying the level of bargaining power (defined as the ability to withhold access to the market) across treatments, we are able to get a clean picture of the behavioral incentives behind arriving at optimal SBT contracts.In order to demonstrate the policy relevance of our findings, we will also conduct a stochastic simulation with the structural model estimated using the combined bakery supplier/Nielsen data. Because this model contains, at its core, a model of utility-maximizing consumers on one side, and pro.t-maximizing retailers and suppliers on the other, we will be able to calculate welfare outcomes (consumer surplus, producer surplus, and total welfare) under a range of market power assumptions. Further, given that the model is estimated econometrically, this simulation will be conducted under a range of stochastic assumptions regarding consumer behavior, as well as supplier and retailer conduct. In this way, we will be able to quantify, within the context of our focal category, the consequences of implementing SBT contracts that are optimal, and those that are not.

Progress 06/01/17 to 05/31/19

Outputs
Target Audience:During the second year, we published an academicarticle with our findings. We also submitted a second paper from this grant to an academic journal. We are currently revising the paper to send it back for publication consideration at this journal. Moreover, we have completed two additional working papers to be submitted to academic journals and presented our researchat several conferences for academics and industry practitioners. Attendees at these conferences included faculty and graduate students from universities in North America, Europe, and Asia. They also included industry practitioners from North America. Finally, we shared our findings with industrypractitioners as part of a workshop we hosted at Arizona State University. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?This project has provided us the opportunity to attend conferences to learn from other colleagues' research in the area of food waste in grocery retail stores and to share our findings on this subject. It also provided the opportunity to mentor graduate students and post-doctoral scholars on a variety of newtechniques for the analysis of data, including the use of econometric models. How have the results been disseminated to communities of interest?We have had the opportunity to deliver talks at Arizona State University as well as other universities and disseminate ourresultsto academic and practitioners communities that are not usually aware of the challenges that suppliers and retailers face in reducing food waste at grocery stores. We have also had a chance to share the insights we have obtained on this subject with the general public through articles in general interst publications such as magazines and newspapers. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? A resurgence of mergers among food retailers and advances in supply-chain technology have raised concerns regarding retail-buyer market power over suppliers. Scan-based trading (SBT) is a new type of contract in which the supplier retains title until the product is checked-out by the customer. Any loss of product between the delivery and checkout is the responsibility of the supplier, not the retailer. If managed properly, SBT provides many advantages for both suppliers and retailers, but the potential for abuse by retailers is clear. Our research provides a better understanding of whether SBT reflects an abuse of market power, or an opportunity for greater supply-chain efficiency. To that end, we study the design of optimal SBT contracts in a principal agent context that align the incentives of supply chain members. Using field data from a major bakery-products supplier, we empirically evaluate the impact SBT has on shrink across a wide array of retail stores served by the supplier. Our results based on average inventory replenishment amounts at retail stores show that SBT contracts account for 10.20% in shrink relative to other contracts. In addition, we conducted an economic bargaining experiment to determine how the effect of retailer market power over the supplier affects the shrink-share and pricing terms negotiated in SBT contracts. Our results suggest thatwhile powerful retailers do tend to have the upper hand in these negotiations, weak vendors could ameliorate or even overcome the influence of retailer power by offering services as a concession and as a bundle with the products that they supply. In so doing, vendors may improve the value of their offerings in the eyes of retailers. When there is a power imbalance in the retailer-supplier relationship in favor of the retailer, we find that theoutcomefrom these negotiations for a supplier was, on average, 3.6% farther than the best possible outcome available whilethe retailer's outcome was, on average, 3.1% closer to its most favorable outcome. However, by bundlingadditional services, the supplier can improve their outcome from these negotiations by up to6.5%, depending on how skewed the power imbalance is in favor of the retailer. We also estimated a structural model of vertical retailer-supplier relationships under SBT in order to test for differences in bargaining power between SBT and traditional supply contracts. Our results show that a switch to an SBT contract arrangement generates an increase of about 46% in bargaining power for retailers, allowing these firms to extract a higher share of the total margin in their vertical relationships with their suppliers. Below, we expand upon the specific objectives addressed in our research. (1) Develop a conceptual framework in which we modelSBT contracts in the framework of a principal-agent transaction, and derive optimal SBTcontracts that align the incentives of suppliers and retailersand maximize the potentialbenefits of SBT contracts. We completed all of the activities for this goal during the previous reporting period pertaining to a review of the literature and the design of a theoretical framework for the principal-agent transactions in SBT contracts. (2) Complete an empirical analysis of field-experiment data obtained from acooperating bakery supplier and Nielsen, including a detailed summary of the data in orderto reveal any stylized facts regarding the apparent relationship between the use of SBTand inventory shrink, SBT and wholesale prices, or SBT and the frequency of stock-outs,and estimates of a difference-in-difference empirical model in which we formally test therelationship between SBT contracts and shrink-share implied by the principal-agent model.The main activitycompleted under this goal for this period involved the publication of a paper at an academic journal. Thefindings reported in this paper show a substantial effect by SBT contracts in generating shrink for food inventory at retail stores (3) Design and estimate a structural model of vertical retail supply relationshipsunder SBT in which shrink-share and wholesale prices are the key control variables.In thismodel, we again use the bakery / Nielsen field experiment data to identify the relationshipbetween contract form and the relative bargaining power of retailers and the supplier.The main activitiescompleted under this goal for this reporting period involved the gathering of data from Nielsen and the preparation and submission of a paper toan academic journal for publication consideration. Another key outcome for this reporting period is the presentation of this paper at several academic conferences. Based on thefindings reported in this paper, there are major differences in bargaining power in the retailer-supplier relationshipbetween SBT and traditional contracts. Specifically, SBT contract arrangements yield greaterbargaining power for retailers, allowing these firms to extract a higher share of the total margin in their vertical relationships with their suppliers. (4) Develop and implement an economic bargaining experiment between buyingand selling subjects in order to determine the effect of exogenously varying bargaining poweron the shrink-share and pricing terms of SBT contracts. We developed and implemented economic bargaining experiments between buying and selling subjects in order to determine the effect of varying bargaining power on the shrink-share and pricing terms of SBT contracts. We conducted three experiments using student subjects in face-to-face environments as well as anonline experiment. We then codified the data from these experiments and conducted data analyses. We find thatwhen there is a power imbalance in the retailer-supplier relationship in favor of the retailer, the supplier's outcomefrom these negotiationswas, on average, 3.6% farther than the best possible outcome available whilethe retailer's outcome was, on average, 3.1% closer to its most favorable outcome. However,by bundlingadditional services, the supplier can improve their outcome from these negotiations by up to6.5%, (5) Simulate the structural model estimated in (3) in order to determinethe impact of SBT contracts, under various shrink and pricing terms, on welfare outcomesfor producer/suppliers, retailers, consumers, and society more generally.Our findings from this simulation produced several new insights into how SBT can work to support the objectives of various stakeholders. These findings are included in a working paper that we wrote and submitted for publication consideration at an academic journal during this reporting period. In particular, we find that there is a spillover effect among retailers already operating under SBT contracts when other competing retailers switched to SBT contracts. Thesespillover effects show that these retailers gained an average of 31% in bargaining power.Moreover, the simulation shows an average total margin increase of 29:08% with retailers switching to SBT contracts earning a larger increase in margin-sharedue largely to a substantial increase in retailer bargaining power (an average of 46.2%) (6) Draw implications of our findings for how SBT contracts can be used toadvantage by both retailers and suppliers, including the potential to minimize food wastethroughout the supply system for our focal category. By identifying the potential contribution of a contract form that is growing in importance, SBT, on the amount of food that is wasted at the retail level, our research highlights an additional mechanism for food loss in the supply chain. Finding empirical evidence that shows the importance of the moral hazard problem in food loss suggests ways in which retail food loss may be mitigated. That is, if both retailers and suppliers better understand the underlying mechanisms of loss, they will enter into negotiations with a more complete understanding of how to mitigate the potentially-negative side effects of their contract.

Publications

• Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Lim, S.; Richards, T.; Choi, M.*; Rabinovich, E. Scan-Based Trading and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. Decision Sciences Institute Annual Meeting, Chicago, IL, November 19, 2018.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2019 Citation: Lim, S.; Richards, T.; Rabinovich, E.; Choi, M*. Scan-Based Trading and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. POMS Annual Meeting, Washington, D.C., May 5, 2019.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2019 Citation: Lim, S.; Richards, T.; Rabinovich, E.; Choi, M*. Scan-Based Trading and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. INFORMS Annual Meeting, Seattle, WA, October 20, 2019.
• Type: Journal Articles Status: Awaiting Publication Year Published: 2019 Citation: Choi, M.*; Rabinovich, E.; Richards, T. Supply Chain Contracts and Inventory Shrinkage: An Empirical Analysis in the Grocery Retailing Industry Decision Sciences. Forthcoming, https://doi.org/10.1111/deci.12356.
• Type: Journal Articles Status: Under Review Year Published: 2019 Citation: Lim, S.; Richards, T.; Rabinovich, E; Choi, M.* Scan-Based Trading and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. W. P. Carey School of Business, Arizona State University.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Lim, S.; Richards, T.; Choi, M.*; Rabinovich, E Scan-Based Trading and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. INFORMS Annual Meeting, Phoenix, AZ, November 7, 2018.
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Choi, M., Grebitus, C., Rabinovich, E., and Richards, T. (2018): Scan-based Trading: An Opportunity for the Food Industry? Evidence from a Negotiation Experiment. Agricultural & Applied Economics Association Congress, 08/05-08/07, Washington, DC.
• Type: Other Status: Other Year Published: 2019 Citation: Grebitus, C. and Richards, T. Can risk preferences and personality explain bargaining behavior? A laboratory experiment for scan-based trading. W. P. Carey School of Business, Arizona State University.
• Type: Other Status: Other Year Published: 2019 Citation: Choi, M., Grebitus, C., Rabinovich, E., and Richards, T. (2018)The Impact Of Bargaining Power And The Mitigating Role Of Service Concession. W. P. Carey School of Business, Arizona State University.

Progress 06/01/17 to 05/31/18

Outputs
Target Audience:During the first year, we had the opportunity to present our findings at different academic conferences. The audience at these conferences comprises faculty and graduate students from universities in the United States. We also had the opportunity to present to practitioners inthe grocery retailing industry some of our findings via one-on-one conference calls. Furthermore, we have been able to disseminate our findings through articles in practice-based publications. Finally, we incorporated the bargaining experimentsinto instructional curricula for undergraduate students at the university's school of business. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?The efforts to reach the communities interested in the work that we are doing mainly involved conference presentations What do you plan to do during the next reporting period to accomplish the goals?Our focus next year will be to complete goals 4, 5, and 6 listed above. For goal 4, our efforts will involve the refinement ofthe brgaining the collection ofadditional data next year. Oncewe complete the data collection, we willcarry out the analyses and generate the results for this We expect to have results available to share next year. For goal 5, we plan to complete the simulations and analyses to determine the impact of SBT contracts, under various shrink and pricing terms, on welfare outcomes for producer / suppliers, retailers, consumers, and society more generally. Upon completion of goals 4 and 5, we will fulfill goal 6,

Impacts
What was accomplished under these goals? Our overarchingobjective is toevaluate how different contract forms, such as vendor managed inventory (VMI) and scan based trading (SBT), between food suppliers and retailers contribute to generate unintended problems in the relationships between these parties. We anticipate that individuals affiliated to food manufacturers and wholesalers as well as grocery retailers will have the most immediate interest in our work. Other stakeholders involved in research and policy-making may also benefit from our work.Below is a description of our progress on each of our specific project goals: 1. We have completed the development of a conceptual framework to model different forms of contracts in the framework of a principal-agent transaction, and derive optimal SBT contracts that align the incentives of suppliers and retailers, and maximize the potential benefits of SBT contracts 2. We have completed an empirical analysis of the field-experiment data obtained from a cooperating bakery supplier and Nielsen. This includes a detailed summary of the data in order to evaluate the relationship between the use of different contracts (SBT and VMI) and inventory shrink. Our results show that the use of SBT contracts generates inventory shrink at retail stores attributable to agency problems in the relationship between vendors and retailers 3. We have designed and estimated a structural model of vertical retail supply relationships under different contract forms to identify the relationship between contract form and the relative bargaining power of retailers and the supplier. We observe that the implementation of SBT contracts between vendors and retailers increases the retailers' bargaining power over the vendors 4. We have developed and implemented an economic bargaining experiment between buying and selling subjects in order to determine the effect of exogenously varying bargaining power on the shrink-share and pricing terms of contract forms involving scan based trading and vendor managed inventory. We collected data from this implementation and are in the process of analyzing. We intend to refine the experiment and gather additional data next year. We expect to have results available to share next year. 5. Another goal is to simulate the structural model estimated in (3) in order to determine the impact of SBT contracts, under various shrink and pricing terms, on welfare outcomes for producer / suppliers, retailers, consumers, and society more generally. This work is ongoing and we expect to complete it next year. 6. Based on the completion of the previous goals next year, we will draw implications of our findings for how SBT contracts can be used to advantage by both retailers and suppliers, including the potential to minimize food waste throughout the supply system for our focal category.

Publications

• Type: Other Status: Under Review Year Published: 2017 Citation: Choi, M.*; Rabinovich, E.; Richards, T. Inventory Shrinkage under Scan-Based-Trading Versus Vendor-Managed-Inventory Contracts: Empirical Analyses in Grocery Retailing.
• Type: Conference Papers and Presentations Status: Other Year Published: 2018 Citation: Lim, S.; Choi, M.; Richard, T.; Rabinovich, E. Scan-based Trading Contract and Bargaining Power: A Structural Model of Vertical Retail Supply Relationships. POMS Annual Meeting Houston, TX, May 6, 2018.
• Type: Conference Papers and Presentations Status: Other Year Published: 2017 Citation: Choi, M.*; Rabinovich, E. The Impact of Shelf Utilization on Inventory Shrinkage: The Moderating Role of Contract Choice. Decision Sciences Institute Annual Meeting, Washington, DC, November 20, 2017.