A Circular Economy (CE) is a system in which products are collected after customers have used them, remanufactured, and then put back into circulation for further use. This reduces the need for virgin materials to manufacture new units and wastage from used units. However, it is well known that delays in inventory returns often lead managers to make suboptimal ordering decisions. This problem is particularly significant in CE systems, where inadequate inventory management can significantly reduce or even negate the benefits of CE. The existing literature lacks evidence on the magnitude of this issue and its key contributing factors. Addressing this gap is essential for developing operational details and maximizing the benefits of CE systems. Using a novel complexity ordering based on a formal characterization of the reverse flow of goods, we investigate the efficiency of managerial decision-making. In our experimental study, participants made inventory replenishment decisions in various CE systems that differed in the complexity of reverse flow. We found that participants' operating costs were substantially higher than the simple benchmark costs, and as the CE system became more complex in reverse flow, the suboptimality of managerial decision-making worsened. Our experimental results provide prescriptions and guidance for practitioners by identifying: (i) CE configurations for which managerial decision-making tends to be better, and (ii) behavioral nudges to improve the efficiency of decision-making.

Managers that consider a closed-loop supply chain just another environmental initiative need to update their thinking. Modern firms that use closed-loop supply chains as a competitive strategy receive many benefits—particularly higher profitability and control over a product's entire lifecycle. In fact, the market for multiple lifecycle products continues to grow, with estimates holding that remanufactured product sales exceed $100 billion per year. As a result of analyzing the ever-growing remanufacturing sector through years of working with managers in numerous industries, various levers and themes surrounding effective closed-loop supply chain strategies became apparent. This chapter presents these findings and shows how firms in multiple industries experienced both successes and failures of their closed-loop supply chain strategies, recently also synonymously called circular economy strategies.

We experimentally investigate whether mass customization enhances sustainability and firm outcomes in a fast fashion context. Fast fashion delivers fashion trends to consumers quickly and cheaply but has detrimental effects on the environment (e.g., waste accumulation, water pollution). To mitigate these harmful effects, we examine how different points of customer involvement in mass customization affect the anticipated number of months to product disposal and willingness-to-pay for mass-customized products. We employ a series of experiments and find that consumer perceptions of the degree of customization increase as the point of customer involvement shifts upstream from Use to Assembly to Fabrication to Design and that the anticipated number of months to disposal and willingness-to-pay increase as the point of customer involvement shifts upstream to Design. We also find that the implementation of customer involvement in mass customization matters. Overall, these results provide evidence that mass customization via Design may not only help slow fast fashion down, which has major sustainability implications, but it may also present a win-win opportunity for both the environment and firms (in terms of the bottom line-provided, of course, that it does not have any major cost disadvantages).

The modeling-based case study is useful for two purposes: introduce closed-loop supply chains and highlight and model some of its unique aspects that the traditional newsvendor formulation does not capture. The case focuses on a third-party remanufacturer (3PR) who buys used cellphones in different quality grades in anticipation of demand. Phones in high grade have been used gently—they have a high acquisition cost but low remanufacturing cost. Low-grade phones have been used extensively—they are cheaper to acquire but have a higher remanufacturing cost. Medium-grade phones have intermediate acquisition and remanufacturing costs. The 3PR needs to trade off these two costs and determine which grade(s) of used phones to buy. The 3PR restores all phones to the same like-new standard during remanufacturing. Extensive use of the case in supply chain management courses shows that in the absence of a mathematical model, students systematically deviate from the optimal decisions because of contextual features. Overall, students believed the case was challenging and that it provides a valuable learning experience, both as an exposure to the closed-loop supply chain domain as well as developing models with industry-specific factors. Supplemental Material: The Teaching Note and Excel solution are available at https://doi.org/10.1287/ited.2021.0254ca .

Customers of many original equipment manufacturers (OEMs) in business‐to‐business markets today demand product servicizing in which instead of buying the product from the OEM they buy the use of a product during a lease. During the lease, (i) the customer uses the product and returns it to the OEM after the use reaches a specific level, (ii) the OEM remanufactures the product in a costly process and sends it back to the customer, and this usecycle is repeated multiple times. The lease terms typically involve a use‐based payment. While a greater product use by a customer brings a larger revenue to the OEM, it also increases the remanufacturing cost incurred by the OEM. We investigate this trade‐off using an analytical model for a contract that is used extensively in industry. For the case of homogeneous customers with a common use rate of the product, we optimize lease payment terms and identify market and product‐remanufacturing characteristics for which the OEM should servicize the product instead of selling it. We show that an OEM should not servicize a product when the customers' use rate exceeds a threshold. This is because, beyond this threshold, the remanufacturing cost increases disproportionably, exceeding the higher usage‐based revenue. Subsequently, we consider a market with two segments with different use rates. We consider two servicizing modes: (i) servicize both market segments or (ii) selectively servicize only one segment and sell the product to the other segment, and the default mode of selling in both segments. We develop optimal lease payment terms for these use‐based servicizing modes, identify thresholds of product and market characteristics for the optimality of these modes. Finally, we extend the results to a market with more than two segments and compare the environmental impacts of the servicizing or sell decision. Numerical results informed by empirical data show that the OEM's loss of profit from choosing a suboptimal servicizing/sell decision can be significant.

OEMs of capital-intensive products are increasingly servicizing their business model, that is, they lease the product to the customer while selling the service. This has led to the forward and reverse flow of products and information between the OEMs and customers. The standard practices and processes of the traditional sell-a-product model do not apply to the servicizing business model. We draw upon our experience with two servicizing OEMs and highlight the inter-functional coordination and development of new business processes needed for OEMs to successfully manage the transition from selling the product to selling the service. Specifically, we discuss the interconnected nature of the business strategy, product development, supply chain, and accounting and finance functions in managing a servicizing business model. Along with close coordination in the strategic dimensions, OEMs also need to make changes in their operational processes to manage the return flow of multiple usecycle products to profitably remanufacture them and re-lease to customers. When operational changes are implemented correctly, the servicizing model also offers several benefits that are not present in the traditional sell-a-product business model.

We investigate the impact of using a clear scoring rule in a sealed bid multi-dimensional (A+B) procurement auction, as frequently used in government procurement. The central procurement agency in Chile (ChileCompra) asked for help to understand how concealing the scoring rule affected buyers. Using an experiment, we analyze the effect of transparently communicating the scoring rule on bidding outcomes by comparing the buyer’s surplus and supplier profits when buyers expressly communicate the weight they place on a nonmonetary (B) attribute, versus when this information is concealed from bidders. In addition, we compare outcomes where the scoring rule is made visible only after the offers are submitted. If the scoring rule is not disclosed, outcomes are poorer for buyers, and sellers see their profits increase.

We investigate the tension between accounting and financial functions of an OEM who manufactures and remanufactures capital‐intensive products. The OEM operates a servicizing business model in which it leases a product, remanufactures the product returned at the end of the lease, and re‐leases the remanufactured product ( usecycle), with multiple such usecycles. The accounting function at the firm would suggest that the OEM: (i) recovers the initial manufacturing cost of the product over its useful life as prescribed by IRS guidelines via depreciation; (ii) remanufactures the product as long as its initial manufacturing cost is not fully recovered; and (iii) removes the product from the books immediately after it has completely depreciated. In contrast, the financial function would consider the initial manufacturing cost of the product to be a sunk cost and would suggest that the firm continues to remanufacture the product as long as the remanufacturing cost is lower than the cost of a new unit. In this study, we investigate and prescribe ways to resolve this tension using analytical models. Specifically, we first show that the optimal durations over which the product should be used and lease prices that maximize the OEM's accounting profit and financial profit are different. We then discuss two operational approaches to align the accounting and operational perspectives while following IRS guidelines. Finally, informed by this analysis, we develop policy prescriptions to promote remanufacturing of multiple usecycle products. Overall, this analysis highlights the role of regulatory guidelines on remanufacturing operations and the inter‐functional effort required by OEMs to successfully integrate remanufacturing operations in their business strategy.

This paper analyzes the business‐to‐business transactions in which a supplier sells assortments of used products to third‐party remanufacturers. The supplier offers used products in different quality conditions, called grades. We model this buyer–supplier transaction as a Stackelberg game in which the buyer chooses his optimal purchase quantity of various grades, and the supplier chooses the optimal assortment and the prices of the grades in the assortment anticipating buyer’s behavior. We first develop an analytically tractable solution to the buyer’s and supplier’s problems. Subsequently, we show several structural properties of the optimal assortment offered by the supplier, including (i) the optimal prices set by the supplier are such that high quality grades have a higher profit margin for the buyer; and (ii) the grades in the optimal assortment constitute a convex hull of the remanufacturing and acquisition costs. We also extend the results to the case when the supplier’s acquisition costs are marginally increasing in the quantity acquired.

Firms are increasingly held accountable for their suppliers' transgressions. Consequently, firms need to develop upstream visibility to exercise control over their supply chains. An emerging body of work has recognized the importance of supply chain visibility and has examined it using analytical models, behavioral methods, and case studies. Still, large‐sample empirical evidence on the benefits of supply chain visibility remains elusive. We seek to bridge this gap by examining conflict minerals disclosures (mandated by the 2010 Dodd‐Frank Wall Street Reform and Consumer Protection Act) and financial reports to evaluate whether firms with greater visibility into their conflict minerals supply chains achieve improved operating and market performance. We use the data from conflict minerals disclosures (Form SD) to distinguish between firms that have high or low visibility into their conflict minerals sources. Then, we use event study methods to analyze differences in operating and market performance between firms with high visibility and firms with low visibility. We find that firms with high visibility into their conflict minerals supply chains achieve higher profitability than comparable firms with less visibility. In addition, we find that firms with high visibility into their conflict minerals supply chains realize improved sales performance and stock market valuations. Our results are relevant to managers because they show that firms can attain operational and market benefits by improving visibility in their supply chains.

Some manufacturers demonstrate their products so that customers can gain experience before making a purchase. We present a novel application of a closed-loop supply chain where product returns from demonstrations of high-end IT equipment are substantial and the major delay in the system is due to the long demonstration time at the client sites. In addition, the product lifecycle is short and the value erodes rapidly over time, with steep drops in the resale revenue when new product generations are introduced. We present a finite lifecycle model that captures the key trade-offs in this environment, that is, either to reuse a collected ex-demo product for a next demonstration or to salvage its residual value in the secondary market and use a new product to satisfy the next demo request. We derive two cost/revenue signals that enable us to distinguish between fast and slow value erosion. We show that the fast/slow erosion decision is dynamic and depends on the rate of value erosion and the length of the demonstration time. We analyze the optimal demo pool strategies and show that in the case of fast erosion it may be better to postpone reuse activities until later in the lifecycle. We illustrate our model using empirical data from a large IT manufacturer and formulate several guidelines so as to better manage high value ex-demonstration product returns. (C) 2019 Elsevier B.V. All rights reserved.

This manuscript defines a typology of remanufacturing based on multiple decades of direct observations across various remanufacturing industries. The manuscript also details how managers adapt their remanufacturing operations and strategies to the idiosyncrasies of the varied remanufacturing industries. The typology identifies four distinct typological groupings based on the dimensions of a firm’s strategic focus and product design philosophy. Before delving into typology and implications on strategic and design issues, the manuscript provides recent information on the current state of the remanufacturing industry based on governmental and industry reports. To assist readers who may be less familiar with the remanufacturing industry and closed-loop supply chains, the discussion also provides a brief overview of remanufacturing processes and the overall remanufacturing industry.

We discuss the optimal raw material acquisition strategy for a third party remanufacturer (3PR). We specifically investigate whether a 3PR should acquire used products or cores in bulk with uncertain quality levels, or in sorted grades with known quality levels; and whether to acquire and remanufacture cores before the demand is realized (planned acquisition), or after the demand is realized (reactive acquisition), or on both occasions (sequential acquisition). When only sorted cores are acquired, we find that, (i) it is optimal to acquire cores in multiple grades to balance acquisition and remanufacturing costs; (ii) if reactive acquisition is possible, it reduces the assortment size (number of grades in which cores are acquired) and the total inventory acquired in the planned acquisition; and (iii) the optimal portfolio of grades to acquire and the optimal acquisition and remanufacturing quantities of these grades can be determined analytically. When bulk cores are acquired in addition to sorted cores, the property of reduction in assortment size of the planned acquisition is preserved. We also show that the 3PR should acquire only a fraction of the demand in planned acquisition, and leave the rest for reactive acquisition. This fraction changes during the lifecycle of a remanufactured product. Using a combination of empirical and realistic data from a smartphone remanufacturer we show that sequential acquisition increases expected profit by up to 8% and 27% over only planned and only reactive acquisitions respectively, and reduces the inventory acquired by up to 21% over only planned acquisition.

The JOM Forum is a new section of the journal that showcases invited essays from influential and experienced scholars from both within and outside Operations Management. The idea is to publish thought-provoking papers that both take stock of past research and look to the future, challenging the way we think of the discipline and our research. The papers published in the Forum are meant to be unique in their style and approach. To this end, we want to give more discretion and active voice to the individuals who write them and allow more freedom to reflect the authors' own style, thinking, and rhetoric. Above all, Forum papers are meant to be provocative so as to lead us out of the box of conventional wisdom. At the same time, they are meant to be empirically grounded and rigorous, just like everything else we publish in the journal. The Forum contributions are by invitation only. Authors are chosen by the editors based on research records, not only to Operations Management but to other disciplines as well. Contributions are peer-reviewed just like all manuscripts submitted to us, but authors are given more freedom to engage in speculation, explore new ground, introduce new empirical contexts, and make connections to other disciplines. As far as the Forum is concerned, all boundaries are meant to be broken. To use Karl Weick's words, an ideal Forum piece is a unique display of disciplined imagination. We hope you enjoy this addition to the journal!

A lot has been written about what constitutes good reviewing and editing, and the guidelines on how to prepare a manuscript for publication are equally abundant. But we feel it is equally important to clarify and discuss the foundations of what constitutes proper author etiquette. There are also many ethical questions related to authoring manuscripts that should be explicated, if only to avoid confusion. Again, the impetus for writing this editorial is empirical: unfortunately, we run into problems and misunderstandings sufficiently often to warrant some reflection and guidance. Here are seven things all JOM authors should keep in mind as they think of their professional duties as authors. The best way to be open about these issues toward the editors is to include with every submission a detailed cover letter. We receive cover letters with less than 50 percent of the submissions—we would like this percentage to be a hundred. In your cover letter, do not just tell us why your manuscript is important and why it makes a contribution, take it also as an opportunity to provide full disclosure and tell us everything we need to know about the manuscript. We also expect our editorial team to follow some simple rules to prevent conflicts of interest. In our experience, by far the most common form of plagiarism is self-plagiarism, which is re-using material one has used in previous publications. We strongly recommend all authors to pay close attention to this. There are no hard and fast rules on what constitutes plagiarism, but you should know that all manuscripts submitted to JOM are submitted to CrossCheck (http://www.ithenticate.com/products/crosscheck), a software that detects plagiarism. The CrossCheck output tells the Editor-in-Chief what percentage of the text in the manuscript can be found in other published sources. Typically, for legitimate manuscripts, this percentage is less than five percent. If the percentage is ten and above, we usually take a closer look at what is causing this; anything over twenty percent is always a cause for concern, and we will contact, without exception, the author and ask for clarification. Sometimes the reason for the high percentage is legitimate and innocuous, for example, it is due to the posting of a working paper version of the manuscript on a web site. This problem can be solved by removing the working paper from the site. In the absence of a legitimate explanation, we typically desk reject the manuscript, because all manuscripts submitted to JOM must be original contributions not published in other sources. It is up to the discretion of the Editor-in-Chief to decide whether a revised version of the manuscript can be submitted. The other form of plagiarism—passing someone else's work as one's own—is obviously a huge cause for concern when it happens and we take it very seriously, because it amounts to intellectual theft. However, this form of plagiarism is so rare at JOM we do not feel it needs to be addressed here. We all know that collecting good data takes time and effort, and getting more than one paper out of a database is tempting. This is fine, but authors must be very clear about this in their submissions. Here is our simple rule of thumb: the data are primary data if and only if they were collected specifically for the purpose of the submitted manuscript. If they were not, the authors must disclose this in a cover letter. Whenever no declaration is made, the editors and the reviewers will assume the data are primary data. If it turns out later in the review process the data are secondary, the Editor-in-Chief may move to reject the paper. Again, using secondary data is not a problem—pretending as if secondary data were primary is. Disclosure is particularly important if the data were collected or the dataset constructed by the researchers. If the authors use already published data (e.g., the Compustat database), then it is obvious it is secondary data, and no further clarifications are required. Disclosure about data also encompasses author effort. For example, if you use a survey database where the data were collected in a collective effort by a group of colleagues, we need to know what your specific role in the data collection effort was. Please be as specific as possible in the cover letter about your role in the data collection process. Sometimes reviewers and Associate Editors handling the manuscript should be provided access to raw data. Unless there are legal or other compelling factors that prohibit raw data disclosure, authors should provide their raw data upon request as well. All data submitted to the journal will be treated confidentially. While we hope JOM would be our authors' first choice, some manuscripts that are submitted to us are clearly rejects from other journals. Our preference aside, there is nothing wrong with submitting to JOM a manuscript that was rejected from another journal. But again, transparency is crucial. If you are submitting a rejected manuscript to us, we strongly encourage you to provide full disclosure of its history. We need to know where the paper was submitted, why it was rejected, and what you did in preparation before submitting to JOM to remedy the problems identified in the reviews. If we find out that the reviewers raised pertinent questions and you chose not to address them (resubmitted the essentially same manuscript to JOM without any changes), we will reject the manuscript. We consider such “recycling” of manuscripts both unprofessional and unethical, because it is a tell-tale sign that authors are more interested in getting the paper published than making sure its quality is intact. We really do not want to see “I reviewed this exactly same paper for Journal XYZ just six months ago” in a reviewer report. However, if your manuscript was rejected from another journal because it was out of the scope of the journal, this obviously does not warrant changes before re-submitting to JOM. Obviously, an Editorial Review Board member or an ad hoc reviewer must never agree to review a paper written by a close colleague, and editors (Editors-in-Chief, Department Editors, Associate Editors) must similarly recuse themselves from evaluating the work of colleagues in situations where conflict of interest is possible. We would classify as close colleague your own doctoral students, students on whose doctoral committees you have served, colleagues with whom you have co-authored in the past five years, and colleagues who are at the same university as you are. Importantly, sometimes conflicts of interest extend to authors as well. Most authors do their research work as part of their salaried job and have no conflicts of interest in submitting their manuscripts to JOM. But in addition to our salaries as faculty members, many of us write books, sell software, give training seminars, and do consulting. Now, if the publication of an article in JOM links either directly or indirectly to such sources of extra income, you must declare this in a cover letter at the time of manuscript submission. This does not mean we think you cannot be trusted, instead, what we mean is that the editors and the reviewers need to know. For example, if you promote in your manuscript a specific statistical method and benefit financially from the sale of a software that implements said method, the editors and the reviewers need to know. What is a significant amount of extra income? We use the guidelines of the U.S. National Institute of Health: all remuneration over $5000 per year must be reported. Remuneration encompasses not only salary and consulting fees, but also royalties, honoraria, paid authorship, and capital gains. If you are unsure about whether or not to disclose something, we would rather have you err on the side of disclosing more than what is needed, hence, “if in doubt, spell it out.” We hold ourselves to the same standards as Editors-in-Chief. When you receive a review on a manuscript you submitted and are invited to revise and resubmit your manuscript, it is your responsibility to address all comments you have received from the reviewers and the editors. Ignoring any single point in a review is sufficient grounds for rejection at the re-submission stage, so please pay close attention to this. The best option is to write a point-by-point response to the review and send this along with the revised manuscript. This ensures you have addressed all comments and concerns, and it helps the reviewers and the Associate Editor enormously in navigating the revision. However, addressing reviewer comments does not mean you have to do exactly as the reviewers or the Associate Editor tell you or that you have to modify the manuscript accordingly. Reviewers and editors are your peers, not your superiors. It is perfectly fine to disagree with a review comment, but this disagreement must be detailed and explicit. And just like we require our reviewers and editors to maintain a scholarly, respectful, and constructive communication style, we expect the same from authors in their communication, no matter how vigorously they disagree with a specific point. Everyone who is associated with JOM has had many of their manuscripts rejected; we all know how disappointing and discouraging rejections are. We fully understand the urge to contact the editor and challenge the rejection, ask for clarification and guidance. Please, avoid the temptation: we have already told you in our decision letter all we want to tell you about the decision. We simply do not have the time to offer further personal editorial services to help authors “fix” rejected manuscripts. The reason is very practical: We absolutely love discussion and debate, and in an ideal world, we would happily discuss each rejected manuscript at great length with the authors. In the real world, we serve the journal with little or no relief from our other duties as faculty members. Last year, in addition to the journal restructuring effort, we both handled over three hundred manuscripts, and ended up devoting somewhere around 600 hours of our time into managing the journal. Please be considerate before you ask us to commit the 601st. We do everything we can to treat all authors and all manuscripts with the care and the expertise they deserve. We encourage our reviewers and Associate Editors to use their expertise to write developmental reviews. We ask you to return the courtesy and respect the decision we make. The only situation in which we encourage you to contact us is if there is a material, factual error, in the review you have received. It has to be an unambiguous factual mistake, not a question of interpretation or policy. It should go without saying that we will not enter into a debate about whether the manuscript is or is not within the scope of the journal: if the Editor-in-Chief rules your manuscript to be out of scope, then it is by definition out of scope. Policy is always our prerogative. We feel compelled to point this out, because the vast majority of author challenges are challenges on policy and interpretation, not fact. In the case of a material, factual error, this is what you should do. Create a table with two columns: “Your Factual Claim” and “My Factual Claim.” Then, write in the “Your Factual Claim” column all the factual statements that you found in the review that you think were not only incorrect but also pertinent to the decision to reject the manuscript. In the “My Factual Claim” column, write what you think is the correct factual claim. Then, e-mail this table along with a cover letter to both Editors-in-Chief. We promise to look at your claim and determine whether corrective measures should be taken. Please keep in mind that nobody serving in an editorial role in the JOM Team has any incentive to reject good papers. Just the opposite, the more high-quality papers we publish, the better for the JOM community. Like we have mentioned in earlier editorials as well as various talks and conference meetings, we live in a digital world and are not page constrained. Accepting one author's manuscript for publication does not “take a slot away” from other authors. There are no slots. Whenever you send a manuscript to us and it is sent out to review, an Editor-in-Chief, a Department Editor, an Associate Editor, and two reviewers will take time off their busy schedule to give attention to your work. If you submit manuscripts to JOM, we expect you to volunteer your time as a reviewer as well. It should not come as a surprise that the demand for reviews exceeds the supply. Consequently, from now on, we have decided to enforce this as editorial policy: the EIC reserves the right to desk reject a manuscript if none of the authors on the manuscript have volunteered their time to review for the journal or any one of the authors consistently declines review requests sent to them from the journal. If you decide to submit manuscripts to JOM, you cannot simultaneously decline requests to review your colleagues' work. Declining to review due to a busy teaching schedule or vacation is no longer considered legitimate. What would you think if we desk rejected your manuscript, saying we are too busy to look at it because we are on vacation? Well, this is exactly how we feel when reviewers decline our review requests. JOM is a scholarly community that operates solely on the basis of reciprocity. We also ask you to be considerate in terms of how many manuscripts with your name on them are in the review process at any given time. We would like this number to be one most of the time, occasionally it may be two. Three is already too many. All these seven points echo the same sentiment: we want JOM to be a scholarly community where the top OM scientists in the world challenge and support one another to produce the highest quality research possible, and where all scholars also volunteer their time to review the work of their peers. Our primary job as Editors-in-Chief is to facilitate unfettered academic debate, the fuel and the fire of scientific progress.

We study a supply planning problem in a manufacturing system with two stages. The first stage is a remanufacturer that supplies two closely-related components to the second (manufacturing) stage, which uses each component as the basis for its respective product. The used products are recovered from the market by a third-party logistic provider through an established reverse logistics network. The remanufacturer may satisfy the manufacturer’s demand either by purchasing new components or by remanufacturing components recovered from the returned used products. The remanufacturer’s costs arise from product recovery, remanufacturing components, purchasing original components, holding inventories of recovered products and remanufactured components, production setups (at the first stage and at each component changeover), disposal of recovered products that are not remanufactured, and coordinating the supply modes. The objective is to develop optimal production plans for different production strategies. These strategies are differentiated by whether inventories of recovered products or remanufactured components are carried, and by whether the order in which retailers are served during the planning horizon may be resequenced. We devise production policies that minimize the total cost at the remanufacturer by specifying the quantity of components to be remanufactured, the quantity of new components to be purchased from suppliers, and the quantity of recovered used products that must be disposed. The effects of production capacity are also explored. A comprehensive computational study provides insights into this closed-loop supply chain for those strategies that are shown to be NP-hard.

Though product reuse through closed-loop supply chains has many benefits for firms, as outlined throughout this book, consumers may not fully appreciate the benefits of buying previously used products. This conjecture led to a series of studies related to how consumers perceive reused products produced in a closed-loop supply chain. Specifically, this chapter summarizes the results from a series of studies that examined how consumers perceive remanufactured and refurbished products. The studies ranged from measuring simple reactions to remanufactured products through experimental manipulation of discount levels and brand equity as a means to determine the appeal of remanufactured products in the general U.S. consumer market. The findings breakdown into multiple levers that prompt consumer interest in remanufactured products including the usually assumed consumer greenness, quality perceptions, discounts, and brand equity. However, the studies also revealed the issue of aversion toward remanufactured products through both disgust and a segment of consumers who only desire new products.

The management of remanufacturing inventory system is often challenged by mismatched supply (i.e., returned units, called cores) and demand. Typically, the demand for remanufactured units is high and exceeds the supply early in a product's lifetime, and drops below the supply late in the lifetime. This supply–demand imbalance motivates us to study a switching strategy to facilitate the decision‐making process. This strategy deploys a push mode at the early stage of a product's lifetime, which remanufactures scarce cores to stock to responsively satisfy the high demand, and switches to a pull mode as the product approaches obsolescence to accurately match the low demand with supply. In addition, the strategy further simplifies the decision‐making process by ignoring the impact of leftover cores at the end of each decision period. We show that the optimal policy of the switching strategy possesses a simple, multi‐dimensional base‐stock structure, which aims to remanufacture units from the i best‐quality categories up to the ith state‐independent base‐stock level. An extensive numerical study shows that the switching strategy delivers close‐to‐optimal and robust performance: the strategy only incurs an average profit loss of 1.21% and a maximum of 2.27%, compared with the optimal one. The numerical study also shows when a pure push or pull strategy, a special case of the switching strategy, delivers good performance. The study offers the managerial insight that firms can use simple, easy‐to‐implement strategies to efficiently manage the remanufacturing inventory system.

There are many things in which Operations Management (OM) researchers can take pride. Since the inception of empirical OM, we have rigorously incorporated measurement reliability and validity into our analyses. In many respects, the OM literature is a few steps ahead of its sister disciplines — incorporating measurement error into analyses is perhaps the best example. We have also made considerable progress in terms of theory development, whether by way of case research or purely conceptual and theoretical analysis. Finally, recent developments in the area of problem solving and design science demonstrate that OM scholars are genuinely interested in solving actual managerial problems and remaining practically relevant. These are all reasons to celebrate the progress in empirical OM. But there are a number of blind spots, many of which continue to be reasons for rejections in the manuscript review process. The purpose of this editorial is to describe some of these issues. Specifically, there are a number of misunderstandings about some of the key methods used in manuscripts submitted to us. There are also some outdated practices that we want to discourage authors from using in their manuscripts. These issues are discussed in this editorial, in a roughly descending order of importance. We all know correlation does not establish causality. It is high time we do something about this. We constantly get manuscripts — based on cross-sectional surveys in particular — where the authors make causal claims. We no longer send to the review process manuscripts that uncritically interpret a cross-sectional correlation of X and Y as support of a causal claim, or more mildly, that the variance of X is driving the variance of Y. This applies to both econometric and structural equation models. The problem with assuming that the variance of X drives the variance of Y is well documented. Ignoring the problem often results in over-permissive tests of substantive hypotheses: we see evidence for our hypotheses even when there is not any. We now require all authors to take steps — theoretical or empirical, preferably both — to address the problem of endogeneity. This is now a standard practice in most top-tier management journals, and it is time for JOM, as a premier operations management journal, to follow suit. The literature on endogeneity is massive, going back almost a hundred years. Roberts and Whited (2013) offer a comprehensive summary of the key issues in the context of corporate finance research. All the issues discussed are directly applicable to OM research as well. In a nutshell, the problem of endogeneity is this: when a researcher is using non-experimental data to test the hypothesis that X has an effect on Y, it is possible that the variance of X is not exogenous but endogenous to the model. The end result is that the model is misspecified. This in fact applies not just to cross-sectional but even longitudinal research. Even if X is measured at t-1 and Y at t, there could be an unobserved variable Z that affects X and t-1 and Y at t. In a recent manuscript submitted to us, authors hypothesized that organizational integration drives employee commitment. Integration was assumed exogenous to commitment. This is a very problematic assumption, because we have many reasons to believe commitment could easily drive integration, making the variance of organizational integration indeed endogenous to the model. The consequence of endogeneity is asymptotic bias in parameter estimation. We must come to terms with the fact that plausible claims about the direction and magnitude of an effect cannot rest on an analysis that completely ignores endogeneity. If our inferences are to be biased, they need to be biased toward being conservative. The problem of endogeneity often has just the opposite effect, it inflates our results. We are not aware of any scientific principles that warrant the use of over-permissive inference. Examination of endogeneity starts with a simple question: What is the source of the variance in the exogenous variables in my model? So far JOM authors have been allowed simply to declare that these sources are exogenous to the model. Authors must take steps toward either demonstrating exogeneity or correcting for endogeneity. Both approaches have the common denominator: they call for addressing assumptions that have thus far gone untested. Endogeneity can probably never be completely eliminated from empirical analysis, and it is well known that many “solutions” create more problems than they solve (Murray, 2006). But there are no good reasons to avoid tackling the issue, at least theoretically. If the problem of endogeneity cannot be addressed empirically by testing for it or using instrumental variables or an experimental research design to mitigate it (Roberts and Whited, 2013), we expect at least a theoretical treatment of the topic in all JOM submissions where the general claim that one variable induces variance in another is made. When arguing that the variance of X gives rise to the variance of Y (causally or otherwise), we expect to see a plausible argument that the direction is indeed from X to Y, not vice versa, or perhaps caused by an omitted variable. Measurement error can also cause an endogeneity problem: if X and Y have a common measurement error source, X will unavoidably correlate with the error term of Y. Finally, sample selection bias may lead to problems very similar to that of endogeneity (Heckman, 1979). While there is definitely a time and a place for cross-sectional research, we strongly encourage cross-sectional researchers to rethink their research designs. We all know how difficult it is to get longitudinal data, but prospective JOM authors must push themselves on this issue and try to fix at least some of the problems of past research by getting out of their comfort zone. If we want to know the magnitude of the effect X has on Y, cross-sectional data is almost guaranteed not to give us a valid estimate. Not only the principles of scientific rigor but also those of practical relevance demand that we get the magnitude right. Many authors continue to build their arguments on the premise that application of statistical inference boils down to rule following. One of the most commonly found “rules of thumb” in manuscripts submitted to us is the claim that a measure is internally consistent if Cronbach's alpha exceeds .70. Nunnally's (1994) book Psychometric Theory is typically cited as the source. But to attribute the rule to Nunnally is a tell-tale sign one has not actually read Nunnally, because if anything, he claimed just the opposite: the criteria for adequate reliability always depend on the context. Lance et al. (2006) unambiguously debunk the “.70 rule.” The technical details of the argument can be found in the works cited in this editorial; there is no need to reproduce them here. If a manuscript submitted to us makes extensive use of unsubstantiated, non-inferential “rules of thumb,” we will desk reject the manuscript. We say non-inferential, because it is crucial to make a distinction between rules that directly link to an inferential test and those that do not. Model fit in structural equation modeling is a good example. Consider two common tools for assessing model fit: the omnibus chi-square test and the Comparative Fit Index (CFI). The chi-square test is an inferential procedure: if the statistic is statistically significant, the model does not fit the data in the sense that the observed and predicted covariance matrices do not match. This is solid inference and methodologically acceptable reasoning. But the claim that a CFI > 0.95 means the model fits the data is not. This is because CFI is a descriptive index, not a test statistic with a commensurate inferential test. A high CFI value simply means the focal model fits the data better than the baseline model. What is the baseline model? It is typically the model where all measured variables are assumed uncorrelated. Using such a baseline model is dubious, because we already know it provides horrible fit for the data. All the CFI thus tells you is that your model fits the data better than a model that does not fit the data at all. It is difficult to see the insight in this conclusion. Lance et al. (2006) discuss the issue in detail, and Tanaka (1993) provides structural equation modelers with a great overview of SEM model fit. H0: the measurement instrument is not internally consistent H1: the measurement instrument is internally consistent Using the “alpha > .70″ rule can help reject the null, but this is not an inferential test, it is merely a social convention that has no methodological basis. What is more, applying the rule misconstrues what methodological texts have actually said. H0: the model fits the data H1: the model does not fit the data The chi-square omnibus test fares much better than the “alpha > .70″ rule. The chi-square test is a valid inferential procedure (it is a test that produces a p-value). A rigorous modeler would consider the fact that the null means the model fits the data, which means low statistical power works to the advantage of the model, not against it. This leads to an over-permissive test, and sometimes this can present a problem. A skillful researcher is able to examine whether or not this is cause for concern. Of course, it is possible to reformulate the null and the alternative hypotheses such that over-permissiveness is not a problem. As far as author requirements, authors of JOM submissions must exhibit an understanding of which rules have a basis in formal statistical inference and which do not. Here, a very simple litmus test works very well: Does the procedure I am using produce a test statistic (with a p-value) or not? At the very minimum, we expect authors to know which rules are simply “urban legends.” This is crucial, because many of the cutoff criteria cited by OM researchers have been thoroughly discredited in the methods literature (Cortina, 2002; Lance, 2011; Lance et al., 2006; Lance and Vandenberg, 2008; Spector and Brannick, 2011). Prospective JOM authors must make themselves aware of this important literature. Citing “an urban legend” will likely lead to desk rejection of the manuscript. Instead of relying on “rules of thumb,” we encourage authors to contextualize their measurement. Indeed, this is what methodological authorities such as Nunnally actually recommend (e.g., Nunnally and Bernstein, 1994, p. 249). Suppose you are interested in estimating a regression model with two explanatory variables (x1 and x2) and a dependent variable (y), and you are assessing measurement reliability. By contextualization we mean asking the question: How does measurement error in my variables affect estimation? It is well known that measurement error in an independent variable is more problematic than in the dependent variable. One can think of measurement error as one of the components of the regression error term, therefore, measurement error in the dependent variable is implicitly already modeled. The statistical consequence of measurement error in the dependent variables is loss of efficiency, which typically does not create problems, particularly if the sample is of reasonable size. Measurement error in the independent variables, in turn, likely causes asymptotic bias to estimates. Although there are no hard and fast rules on the consequence, the resultant bias is roughly proportional to the amount of measurement error (Kennedy, 2008; Maddala, 1988). Increasing sample size does not fix the problem, because bias is asymptotic. How many authors citing the “alpha > .70 rule” for an independent variable realize that they are implicitly admitting that an asymptotic bias of up to 30 percent in a parameter estimate is acceptable? How much sense does it make to report parameter estimates with three significant digits when even the first digit is likely wrong? The Variance Inflation Factor (VIF) to test for multicollinearity is a perfect example of lack of contextualization. There is nothing wrong with the VIF itself, but every recommended cutoff must simply be ignored. In short, the VIF tells the researcher how much the variance of the parameter estimates has been inflated due to collinearity of predictors. But the VIF value has no meaning until one has looked at the magnitude of the variances of the parameter estimates. In large samples, variances of estimates are very small, therefore, even a tenfold (VIF = 10) increase may not present any significant problems. In a small sample, doubling of the variance (VIF = 2) may already be cause for concern. All generally recommended cutoffs that ignore sample size are nonsense. In general, statistics experts (at least sensible ones) never give recommendations without incorporating the context. If you ask an expert on estimation theory which estimator you should use in your model, you will not get an answer until you have described in detail your model, your data, the distributions of your variables, and the extent to which you believe your model is correctly specified. Should SEM researchers start using Bayesian estimators (BSEM), for instance? This is what the architect of one of the most commonly used estimators had to say only a few years ago: “Much more experience is needed… More needs to be learned about the performance of BSEM parameter posterior estimation using different informative priors for different types of models, sample sizes, and variable distributions” (Muthén and Asparouhov, 2012, p. 333). The fundamental problem with the use of various cutoffs for reliability and validity is that they turn measurement questions into yes/no issues, when it should be obvious that most issues that have to do with numbers are matters of degree. It is time to embrace this premise in empirical analysis. We understand that OM scholars are not statisticians, but one needs to understand the tools one uses to an appreciable depth. The most alarming example is the continuing use of Partial Least Squares modeling. We are desk rejecting practically all PLS-based manuscripts, because we have concluded that PLS has been without exception the wrong modeling approach in the kinds of models OM researchers use. Most of the time, use of PLS is (incorrectly) justified by saying that PLS is suitable for small samples, that it should be used when one has formative indicators in a measurement model, or that it is suitable when the Maximum Likelihood estimator fails to converge to a solution. All are poor excuses for using PLS. Claiming that PLS fixes problems or overcomes shortcomings associated with other estimators is an indirect admission that one does not understand PLS. Consequently, we will automatically desk reject a manuscript that makes incorrect claims about the applicability of the estimator (obviously, any estimator, not just PLS). After all, choosing the right estimator is one of the most important steps in statistical inference. The primary prescription is simple: never use an estimator or a modeling method you do not understand. As far as PLS is concerned, there have been lots of discussions in recent years in the methods literature; there is no need to reproduce the technical details here. If you think PLS is the appropriate estimator, take a look at Marcoulides and Chin (2013), Rönkkö and Evermann (2013), and McIntosh et al. (2014). Upon reading these articles, if you are still convinced PLS is a suitable estimator for your model, we welcome your PLS-based analysis to the journal. In your manuscript submission, clearly justify the use of PLS in light of the three articles cited above. This obviously applies to all estimators and modeling techniques. Many statistical programs (such as Stata and Mplus) have at least a dozen estimators from which an author can choose. This choice must be made transparently. The choice is transparent when the author clearly discusses both the strengths and the weaknesses of the chosen estimator. Not a single PLS manuscript submitted to JOM has discussed the weaknesses of the estimator. Authors should always avoid rhetoric such as “expert X has suggested that estimator Y be used.” Such rhetorical appeals must be replaced with methodological justification. Ketokivi and Schroeder (2004) showed that common method bias is impossible to address in survey research in an adequate manner unless one uses multiple informants per observational unit. Yet, many authors of single-informant studies make strong claims that common method bias is not a concern in their study. Such claims are dubious, because the effect of one of the well-known sources of potential bias — the informant (e.g., Campbell and Fiske, 1959; Phillips, 1981) — simply cannot be tested. Authors often use Harman's (1967) single-factor test as the inferential tool to test common method bias. Using the word test to describe the technique is, however, misleading: Harman's test is a more or less arbitrary procedure with no commensurate inferential test. Indeed, Podsakoff et al. (2003, p. 889) note that “despite the fact this procedure is widely used, we do not believe it is a useful remedy to deal with the problem.” It should go without saying that authors must not use procedures that are not useful. You should always be aware of what can and cannot be tested. The conclusion of no common method bias must be made with much caution (if at all) if the statistical test used to examine it is weak. Harman's test is perhaps the best example. For survey researchers, we have a very simple recommendation: either give up single-informant surveys or stop making strong claims about common method bias. If the research design relies on only one informant per observational unit, there is no way to determine what proportion of item variance is trait variance. Further, as Podsakoff et al. (2003) aptly note, all techniques — including the most sophisticated ones — have problems associated with them. Prospective JOM authors must understand what these weaknesses are, and accordingly, not assume there are unambiguous technical fixes to common method bias. There are no straightforward remedies to common method bias, because its sources are diverse and complex; this is the key message in Podsakoff et al. (2003). All we can recommend is that JOM authors use the most rigorous test available with their data. If the most rigorous test is Harman's single-factor test, authors should be extremely cautious with their conclusions. Addressing common method bias must really start at the research design phase: most effective remedy is to be ex ante smart about the issues. Many ex post analyses can only diagnose whether or not there is a problem — if there is a problem, there is usually not much the researcher can do at that point. If you cited Baron and Kenny (1986), Podsakoff and Organ (1986), Harman (1967), and Fornell and Larcker (1981) in your work back in the 1980s, you were probably fine. In 2015, you need to be careful. While statistical theory itself has not progressed all that much, the software applications that we all have on our desktops have massively improved. We have many solutions available to us now that we did not have in the 1980s; many of the shortcuts we took back in the day no longer need to be taken; many of the assumptions we were forced to make can now be relaxed. It behooves us to stay on top of current methodological developments, and accordingly, what was accepted in the journal ten, twenty years ago, is not necessarily acceptable anymore. As a general rule, prospective JOM authors should be cautious with benchmarking for methodology research published more than twenty years ago. Authors must of course give credit where credit is due, but it should go without saying that using Fornell and Larcker (1981) — an article published 35 years ago in a marketing journal — for methodological guidance in an operations management article submitted to a top journal in 2015 should be done with much caution. Older texts have a lot of useful material, but they are not to be used as research manuals. It is no exaggeration to say that new methodological developments come out every month. Most of these developments are minor, but many of them are noteworthy. Therefore, instead of using Baron and Kenny (1986) as a resource on how to deal with mediation hypotheses, one might look at Hayes (2013) and James et al. (2006) for updated approaches. One can even find published SPSS and SAS routines for testing mediation (Preacher and Hayes, 2004). There is just no excuse for not using up-to-date tools. How about a check on Google Scholar to find out whether there have been any new developments in methodology relevant to your work before you submit your manuscript? The Organizational Research Methods journal is also a wonderful resource for all management scholars. For instance, a lot of the relevant discussion on PLS has been published in ORM. The essence of science lies in the collective quest toward continuous progress. In terms of methodology, this means we need to strive for stronger inference. All the criteria described above have this primary objective. Most of us are, in one way or another, interested in the question How important is X as far as the outcome Y is concerned? Moving toward stronger inference will lead us toward an unbiased estimate of this effect. We want to encourage both authors and reviewers not just to become aware of the key problems, but also do something about them. All the problems described in this editorial are remediable, and we are already seeing authors engaging these problems in the more recent submissions to the journal. This is very encouraging.