Adem €Orsdemir, Eda Kemahlıo�glu-Ziya, Ali K. Parlakt€urkKenan-Flagler Business School, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599, USA
W e consider an original equipment manufacturer (OEM) who faces competition from an independent remanufacturer(IR). The OEM decides the quality of the new product, which also determines the quality of the competing remanu-
factured product. The OEM and the IR then competitively determine their production quantities. We explicitly character-ize how the OEM competes with the IR in equilibrium. Specifically, we show that the OEM relies more on quality as astrategic lever when it has a stronger competitive position (determined by the relative cost and value of new and remanu-factured products), and in contrast it relies more heavily on limiting quantity of cores when it has a weaker competitiveposition. The IR’s entry threat as well as its successful entry can decrease the consumer surplus. Furthermore, our resultsillustrate that ignoring the competition or the OEM’s quality choice leads to overestimating benefits of remanufacturingfor consumer and social welfare. In addition, we show an IR with either a sufficiently weak competitive position (so theOEM deters entry) or a sufficiently strong one (so the OEM is forced to limit quantity of cores) is desirable for reducingthe environmental impact. Comparing our results with the benchmark in which the OEM remanufactures suggests thatencouraging IRs to remanufacture in lieu of the OEMs may not benefit the environment. Furthermore, the benchmarkillustrates that making remanufacturing more attractive improves the environmental impact when the remanufacturer isthe OEM, while worsening it when remanufacturing is done by the IR.
Key words: product quality; remanufacturing; competition; environmental impact; social welfareHistory: Received: February 2011; Accepted: November 2012 by Gilvan Souza, after 3 revisions.
1. Introduction
Remanufacturing operations involve taking usedproducts, bringing them back to as-new condition,and selling them again (Atasu et al. 2010). Theseactivities in an industry can be carried out either bythird-party independent remanufacturers (IR) or byoriginal equipment manufacturers (OEM). Especiallyin the United States, the majority of remanufacturingis done by IRs (Hauser and Lund 2008). The samestudy finds that the remanufacturing industry in theUnited States is worth $53 billion, which means thatIRs are not an insignificant competitive threat toOEMs. OEMs try to fend off competition from IRsthrough limiting quantity, specifically by creatingscarcity of cores available for remanufacturing (e.g.,by offering free take-back of cores from consumers[HP 2010a] or making cores ineligible for remanufac-turing [Lexmark 2010] and rarely through litigation[e.g., HP 2010b]).There is also evidence that OEMs change their
product designs with remanufacturing concerns inmind. For example, Subramanian et al. (2011) arguethat HP refrains from using common print heads inits business inkjet printers because doing so makesthe IRs a bigger competitive threat in the market.Atasu and Souza (2011) describe how Xerox and
Kodak take remanufacturing into consideration whenthey design their products. An important productdesign decision that is the focus of this study is qual-ity. Following Moorthy (1988) and Desai (2001), wedefine quality as an attribute that exhibits the “moreis better” property, so given the same price, all cus-tomers prefer the higher quality product. It is wellknown that firms can use quality as a competitivelever; however, in the remanufacturing context thedynamics around the quality decision are intricatebecause when an OEM increases its product quality,it also increases the quality of the remanufacturedproduct to a certain extent. Therefore, the results onproduct quality from studies that consider competi-tion between independent products (e.g., Desai 2001,Moorthy 1988) do not immediately extend to theremanufacturing context. Thus, in this study weexamine how an OEM can use product quality along withquantity as a competitive lever against an IR.Remanufacturing is generally perceived as an envi-
ronmentally friendly end-of-use management optionfor many products. Commonly cited benefits includediversion of discarded products from landfills,reduced virgin raw material usage, and reducedenergy usage when compared to manufacturing (USEnvironmental Protection Agency 1997). At thesame time, Gutowski et al. (2011) find that while
remanufacturing itself uses less energy than manufac-turing, remanufactured products may be less energyefficient. Thus, the relative environmental impacts ofnew and remanufactured products should be care-fully considered, and the total environmental impactof remanufacturing in a given market is not clear.Recently, we have seen a surge of activities that pro-
mote remanufacturing. For example, the AutomobileParts Remanufacturers Association introduced theRecycling/Remanufacturing Tax Credit Bill, HR 5695(The Remanufacturing Institute 2008) and campaignssuch as Manufactured Again (Motor and EquipmentRemanufacturers Association 2011) work to increaseremanufacturing levels by increasing consumerawareness. An underlying tenet of these activities isthat remanufacturing is good for the consumer. How-ever, just like total environmental impact, the socialwelfare implications of remanufacturing, especiallywhen it is conducted by a third-party are not wellunderstood. To this end, we research how the competi-tion between the IR and the OEM affects total environmen-tal impact and social welfare, specifically when the OEMcan adjust product quality in response to competition.We consider an OEM who faces competition from
an IR. The OEM decides the quality of the new prod-uct, which also determines the quality of the compet-ing remanufactured product. The OEM and the IRcompetitively determine their production outputs,which determine the prices of the new and remanu-factured products. The remanufactured product canbe perceived as inferior in quality but cheaper to man-ufacture. We study the relation between the competi-tive positioning of the OEM and the IR and how theOEM chooses to compete with the IR as well as theenvironmental and social welfare implications of thischoice. In our base model, the OEM sells the newproduct and remanufacturing is done solely by theIR. In addition to our base model, we consider severalbenchmarks: a monopolist OEM without remanufac-turing capability (NR benchmark), a monopolist OEMwith remanufacturing capability, and competitionwith exogenous quality decision. These benchmarkshelp tease out the effects of competition, the OEM’squality choice, and the type of the remanufacturingfirm on our results.Even though most remanufacturing in the United
States is done by IRs, OEMs like Xerox, Kodak, andCaterpillar have remanufacturing operations, too. Inan extension to our base model, we study how theanswers to our research questions change when theOEM remanufactures instead of the IR. Comparingour findings with the results of this extension, we areable to provide insights on how the environmentaland social welfare benefits of remanufacturingdepend on the type of company (IR vs. OEM) offeringthe remanufactured product. When faced with com-
petition from an IR, some OEMs like Lexmark chooseto collect cores and dispose of them rather thanremanufacture in-house. We analyze this scenario asan extension to our base model as well and provideinsights regarding when the OEM prefers to collectcores to compete with the IR. We now summarize ourkey findings:
• We explicitly characterize how the OEM com-petes with the IR in equilibrium. When theOEM has a significant competitive advantage(which is determined by the relative cost andthe perceived quality of the remanufacturedproduct vis-à-vis the new product and isexplained in detail in section 3), it deters theIR’s entry by choosing a quality level that ishigher than it would if the IR was not in themarket. In contrast, when the IR has a signifi-cant competitive advantage, the OEM reducesproduction and, hence, decreases the numberof cores the IR can remanufacture. In between,the IR enters the market and does not encoun-ter core shortage. In this region, when theOEM has the competitive advantage, it choosesa higher quality level compared to the NRbenchmark to emphasize its advantage. Whenthe IR has the competitive advantage, theOEM chooses a lower quality level to de-emphasize its competitor’s advantage. Ourresults show that when the OEM has a stron-ger competitive position, it is more likely torely on quality as a strategic lever, whereaswhen the IR’s competitive position gets stron-ger, the OEM is more likely to rely on limitingcore availability.
• The IR’s entry threat as well as its successfulentry can decrease the consumer surplus com-pared to the NR benchmark; that is, remanu-facturing may harm consumer welfare. This isbecause the OEM chooses an inefficiently highquality level to deter or weaken the IR. In con-trast, when the product quality is exogenouslyfixed, the consumer surplus always increaseswith remanufacturing. We show a similarresult for the social surplus. These results arein contrast with our monopoly remanufactur-ing benchmark, which shows remanufacturingby an OEM always benefits the consumer andsocial surplus. There are two factors in playhere: (i) an IR chooses to remanufacture evenwhen the perceived value to cost ratio of theremanufactured product is unfavorable relativeto the new product. In other words, the OEM’sremanufacturing incentives are better alignedwith consumer and social surplus than thatof the IR. (ii) When the OEM remanufactures
itself, it chooses product quality more effi-ciently as far as consumer and social surplusare concerned. Overall, our results illustratethat ignoring competition or the OEM’s qualitychoice lead to overestimating benefits ofremanufacturing for consumer and socialwelfare.
• We also study the environmental impact ofremanufacturing. When the OEM deters theIR’s entry through increasing quality, the envi-ronmental impact always decreases. Basically,a higher quality product implies a smaller salesvolume, reducing the environmental impact.When the IR enters the market and remanufac-tures, the environmental impact decreases ifand only if the remanufactured product has asufficiently smaller per unit relative impactcompared to the new product, and we explic-itly characterize this critical threshold. As faras environmental impact is concerned, an IRwith either a sufficiently weak competitiveposition (so the OEM deters entry) or a suffi-ciently strong one (so the OEM is forced tolimit quantity of cores) is desirable. Whenneither the OEM nor the IR has a strong advan-tage, the bitter competition between the twoincreases the total sales quantity aggravatingthe environmental impact. Comparisons withour NR benchmark show that when remanu-facturing has a competitive advantage deter-mined by its relative cost and perceivedquality, remanufacturing by the OEM is morelikely to reduce environmental impact thanremanufacturing by an IR. This is due to twofactors. (i) Competition increases the salesquantity worsening the environmental impact.(ii) The OEM can choose a lower quality levelwhen competing with the IR, which alsoincreases the sales quantity. Our results canhave important policy implications: encourag-ing OEMs to remanufacture their own productsmay be more beneficial for the environmentthan encouraging IRs to remanufacture.
• For the two alternative models we consider, inwhich the OEM can also remanufacture or itcan preemptively collect cores, we showthrough numerical studies that the way theOEM chooses to compete with the IR is similarto our base model. Consistent with ourinsights from the base model, the OEM followsa deterrent quality strategy when remanufac-turing does not have a strong value proposi-tion; in contrast, it uses a deterrent quantitystrategy (remanufacturing itself or collectingcores and disposing of them) when theIR’s remanufacturing becomes a bigger threat.
Furthermore, we find making remanufacturingmore attractive, by either lower cost or higherquality perception, can worsen the environ-mental impact when remanufacturing is doneby the IR; in contrast it lessens the environ-mental impact when the OEM is remanufactur-ing. Thus, the consequences of these incentiveson environmental impact critically depend onthe type of the remanufacturing firm.
• We demonstrate the robustness of the equilib-rium structure, which shows how the OEMchooses to compete with the IR, under threedifferent extensions: the IR incurs an addi-tional cost independent of the quality level; theperceptual quality gap between new andremanufactured products is independent ofproduct quality; and the OEM and the IR com-pete in prices. Comparison of our results fromthe base model and the extensions, however,shows that the effect of IR’s competitive threaton the OEM’s quality choice may criticallydepend on how the cost and perceived qualityof the remanufactured product are modeled.Furthermore, the implication of remanufactur-ing on the social and consumer surplus andenvironmental impact can be sensitive to thetype of competition (price vs. quantity).
2. Literature Review
The closed-loop supply chain literature has studied anumber of questions that arise when a remanufac-tured product is introduced into the product mix. Theliterature makes different assumptions regardingwho produces the remanufactured product: a monop-olist OEM who also sells the new product (e.g., Esen-duran et al. 2010, Ferrer and Swaminathan 2010), anIR competing with an OEM (e.g., Esenduran et al.2012, Ferrer and Swaminathan 2006, Majumder andGroenevelt 2001), or an OEM who faces competitionfrom another firm (e.g., Atasu et al. 2008, Heese et al.2005). Ferguson and Toktay (2006) compare, from thepoint of view of an OEM, the profitability of introduc-ing a remanufactured product versus collecting anddisposing of used products to deter the entry of an IR.This stream of literature studies how the competitionbetween new and remanufactured products affectsthe pricing and quantity decisions of the OEM (afeature that also exists in our model) but does notcapture the OEM’s endogenous quality decision. Weextend this literature by allowing the OEM to explic-itly set product quality. We characterize how theOEM uses two modes of competition—quality andquantity—as its competitive position vis-à-vis the IRchanges. We also research how the competitionbetween an OEM and an IR and the OEM’s ability to
choose the quality level affect consumer surplus andthe product’s total environmental impact.How competition affects a firm’s quality choice has
been studied extensively in marketing literature. Onefundamental difference of our model is that the OEMmakes the quality decision and its decision locks in thequality of the remanufactured product whereas in theextant literature, competing firms are allowed tochoose their own quality levels independently. Thisdifference leads to significantly different insights. Forexample, Moorthy (1988) shows that in a duopolywhen firms choose their quality levels first and thencompete in prices, consumer surplus is higher than inthe monopoly case. In our model, consumer surplusmay be lower than the monopoly case because theOEM takes advantage of the interdependency betweenthe products and may inefficiently increase or decreasequality in order to weaken the IR’s competitive posi-tion. Desai (2001) also models a duopoly but with sym-metric firms. In contrast, the asymmetry between theOEM and the IR determines their relative competitivepositioning, which plays a key role in our results.In the operations management literature, a number
of studies examine how competition impacts firms’quality decisions or related variables such as servicelevels and waiting times. Banker et al. (1998) modelthe quality and price competition between two manu-facturers. They find that product quality increaseswhen a low-cost entrant enters the market where anincumbent has the intrinsic demand advantage. Wereach the exact opposite conclusion, and this isbecause, in their model, both firms are allowed tochoose their own quality levels independently. Inother work in operations management (e.g., Bernsteinand Federgruen 2004, Tsay and Agrawal 2000) thereis an interdependency between quality and demand/supply parameters, and imbalance between supplyand demand deteriorates quality. In contrast, in ourmodel quality is an intrinsic product attribute inde-pendent of the magnitude of demand.We study competitive quality choice for a remanu-
facturable product. Quality is an important productdesign decision and has been of great interest in thenew product development literature (e.g., Fishmanand Rob 2000, Plambeck and Wang 2009, Souza et al.2004). However, these papers are mainly concernedwith sequential quality improvements whereas qual-ity choice is made only once in our model. Further-more, the remanufacturing context has some uniqueaspects: the remanufactured product’s cost and qual-ity level depend on the new product’s quality leveland the OEM can limit the cores that the IR can accessfor remanufacturing, which adds another layer ofinterdependence. Here, we contribute by studyingquality choice for a product that competes with aninterrelated product.
In the context of product recovery, few studies con-sider product quality explicitly. In Debo et al. (2005)and Robotis et al. (2009), quality refers to the remanu-facturability level of the returned product, whichreduces the remanufacturing cost; this is differentfrom our definition. Debo et al. (2005) model amonopolist OEM and research whether the OEMshould sell a remanufacturable product and if so,what the level of remanufacturability should be. In anextension that allows competition with IRs, they findthat as remanufacturing competition intensifies, theremanufacturability level of the product goes down.However, we find that as the IR becomes more com-petitive up to a threshold level, product quality goesup. Robotis et al. (2009) consider a monopolist andshow that uncertainty in remanufacturing cost maylead to higher reusability investment. Subramanianet al. (2011) study how remanufacturing threat of anIR affects the component commonality decision for anOEM selling two vertically differentiated productswith exogenous qualities.Atasu and Souza (2011) is the closest to our work.
They consider a monopolist who remanufactures in-house and study the effect of three product recoveryforms, that is, quality recovery (remanufacturing is anexample), profitable material recovery, and costlyrecovery, on quality levels. They find that qualityrecovery and costly recovery lead to increased qualityand decreased environmental impact while profitablematerial recovery leads to decreased quality andincreased environmental impact. Furthermore, qual-ity recovery benefits the consumers, but costly recov-ery reduces the consumer surplus. Atasu and Souza’swork clearly demonstrates that not all forms of recov-ery are equally beneficial for the environment and theconsumers. In this work, we confine ourselves to asingle recovery form, that is, remanufacturing, butconsider the competition between the OEM and theIR. We also study how the product quality level andbenefits of remanufacturing depend on the party(OEM or IR) doing the remanufacturing. We find thatwhen an OEM and an independent remanufacturerare in competition, remanufacturing may indeedresult in decreased quality and increased environ-mental impact.
3. Model Overview
We consider an OEM selling a new product and an IRselling the remanufactured product. We begin byintroducing the demand model, discuss the cost struc-ture, and finally describe the firms’ decisions.Each customer considers new product, remanufac-
tured product, and no purchase options and choosesthe one that maximizes her utility. We model con-sumer preferences as in Moorthy (1988). Consumers
are heterogeneous in their willingness to pay for qual-ity and are uniformly distributed over a boundedsupport with unit density, which we normalize to[0,1]. A consumer of type h ∈ [0,1] is willing to pay hsfor a product of quality level s. This implies that,everything else being equal, all consumers preferhigher quality over lower quality. Given that pn is thenew product’s price, the utility a type h customerreceives from the new product is hs� pn. Consumersoften perceive the remanufactured product as beingof inferior quality. We capture this by modeling con-sumers’ willingness to pay for the remanufacturedproduct as a d fraction of the new product whered ∈ (0,1). Consumption of the remanufactured prod-uct provides a utility of dhs� pr where pr is theremanufactured product’s price. This implies that thequality gap between the new and remanufacturedproducts is proportional to product quality s. Amongothers, Ferguson and Toktay (2006) and Atasu et al.(2008) model demand and the relative valuation ofthe remanufactured product similarly. In section 8.3,we consider an alternative model where the qualitygap is independent of product quality.The unit variable cost of producing a new product
with quality level s is bs2 where b is a scaling parame-ter and does not alter our insights (e.g., Atasu andSouza 2011, Desai 2001, Moorthy 1988). The qualitylevel of the new product impacts the remanufacturingcost, too. Since remanufacturing brings a product tolike-new condition by replacing older and worn-outparts, it is costlier to repair and replace the higherquality parts of a higher quality product. At the sametime, remanufacturing a product costs less than man-ufacturing a product because some parts are reused.The remanufacturing cost is proportional to the costof new product, specifically, it is equal to bas2. Assuch, our base model does not consider a quality-independent cost term. An extension in section 8.2allows for such an additional cost term as in Atasuand Souza (2011). Here, a ∈ (0,1) is an indicator of theremanufacturing cost advantage and it decreases asthe cost savings from remanufacturing increases. LikeDebo et al. (2005) and Ferrer and Swaminathan(2006), we assume that the remanufacturing costsubsumes the cost of all remanufacturing-relatedactivities.The order of decisions is as follows. First, the OEM
decides the quality of the product s. Then the OEMand the IR competitively choose new product andremanufactured product quantities, qn and qr, that aresold in a single period. The IR’s remanufacturingquantity is constrained by the available cores, whichis determined by the new product quantity. The IRcan also choose to stay out of the competition bychoosing to remanufacture zero units. Finally, con-sumers make their choices.
We consider a product that has a single (long) life,and the quality decision is made at the beginning ofthis long life cycle. Single period models have previ-ously been used in Atasu and Souza (2011), Agrawal etal. (2011) and Subramanian et al. (2011) in the sustain-able operations management literature. This approach,which focuses on steady-state profits, facilitates analyti-cal tractability in our model and allows us to focus onour research questions. Furthermore, the OEM and IRengage in Cournot type quantity competition in ourmodel as in Ferguson and Toktay (2006), Debo et al.(2005), and Atasu et al. (2009). Both quantity and pricecompetition models are extensively used in the OMliterature. While our base model adopts quantity com-petition, we also study price competition showing thatour equilibrium results propagate.Following Johnson and Myatt (2006), the OEM’s
and the IR’s chosen quantities and customer choiceslead to the following prices for the new and remanu-factured products:
pn ¼ sð1� qn � dqrÞ; pr ¼ dsð1� qn � qrÞ:The above equations assume that the product’s use-
ful lifetime is one period, it can be remanufacturedonly once, and all recovered cores are in good enoughshape for remanufacturing (e.g., Atasu and Souza2011, Debo et al. 2005, Ferguson and Toktay 2006).Our model can be extended such that only a fractionof the cores is available for remanufacturing, but fortractability and to keep the focus on our researchquestion, we only consider the case where all corescan be remanufactured.We derive the equilibrium by backward induction.
For a given quality level s, the OEM and the IR playthe quantity game. Formally, the OEM solves
maxqn
pOEMðqnjsÞ ¼ ½sð1� qn � dqrÞ � bs2�qns.t. qn � 0
and the IR simultaneously solves
maxqr
pIRðqrjsÞ ¼ ½dsð1� qn � qrÞ � abs2�qrs.t. qr � 0; qr � qn:
This solution approach is the same as in Agrawalet al. (2011). The IR’s problem has an additional con-straint reflecting the fact that the remanufacturedproduct quantity cannot exceed the new productquantity, that is, qr � qn, the core availability con-straint. Finally, the OEM chooses the optimal qualitylevel s� by solving maxs pOEMðsjq�nðsÞ; q�r ðsÞÞ. Theresulting equilibrium is described in the next section.Note that we use the superscript (*) to denote equilib-rium values throughout the article.
In addition to our base model, we consider themonopoly no-remanufacturing (NR), monopoly reman-ufacturing, and exogenous quality benchmarks. Themonopoly NR benchmark considers a monopolistOEM who sells only the new product, deciding thequality and quantity of its product. In the monopolyremanufacturing benchmark, a monopolist can sellboth new and remanufactured products. In the exoge-nous quality benchmark, the quality of the new prod-uct is fixed at level sf and the OEM competes with theIR using only quantity. Thus, in this benchmark, whenthe IR enters the market, the OEM adjusts its productquantity but cannot change exogenous product quality.These benchmarks help us characterize the effects ofcompetition, remanufacturing, and OEM’s qualitychoice on our results. The equilibria for the bench-marks and all the proofs are provided in the OnlineSupplement.
4. Equilibrium
In this section, we discuss the decisions of the OEMand the IR in equilibrium. As will be evident, aremanufactured product’s relative cost-to-value ratio,a/d, simply referred to as the cost-to-value ratio, playsan important role in our result. When the a/d ratiodecreases, the IR’s competitiveness increases and viceversa for the OEM. This is because increasing the a/dratio indicates that either the cost of remanufacturinggoes up or the consumer perception of the remanufac-tured product goes down. Specifically, when a/d isgreater than 1, the OEM has the cost-to-value advan-tage against the IR. In contrast, when the cost-to-valueratio is smaller than 1, the IR has the advantage. Con-sider medical imaging equipment and printer car-tridges for two examples that fall on two oppositeends of the spectrum. Remanufacturing medicalimaging equipment (e.g., computer tomography andmagnetic resonance imaging) has a high marginal costdue to high technology components used in theseproducts. In addition, hospitals are skeptical aboutbuying remanufactured imaging devices (Elsberry2002) since they can have a direct impact on patients’health. Thus, medical imaging equipment can becharacterized by a large a/d ratio. In contrast, printercartridges possess a small a/d ratio due to low costand high consumer acceptance of remanufacturedcartridges. In fact, the cartridge industry is one of theprominent examples where the competition betweenthe IRs and the OEMs (e.g., Lexmark, HP) is verysevere. The next proposition describes the equilib-rium.
PROPOSITION 1. The following characterizes the equilib-rium regions. The equilibrium quality and new and reman-ufactured product quantities are provided in Table 1.
R1. If ad � 2, the IR cannot enter the market and the
OEM acts like a monopoly.
R2. If 8�d4þd � a
d \ 2, the IR is a threat and its entry isdeterred by the OEM.
R3. If dð18�8d�2d2þd3Þð4�dÞ2 \ a
d \8�d4þd, the IR enters the market
but does not remanufacture all available cores.
R4. If 0\ ad � dð18�8d�2d2þd3Þ
ð4�dÞ2 , the IR enters and remanu-factures all available cores.
Figure 1 graphically depicts the equilibrium regionsin the proposition. Note that the cost-to-value advan-tage shifts from the OEM to the IR as we move fromregion R1 to R4. In region R1, the IR does not pose athreat due to its severe cost-to-value disadvantageand the OEM acts as a monopolist leading to the sameoutcome as the NR benchmark.In region R2, the IR is a competitive threat. How-
ever, the OEM is able to deter entry by choosing ahigher level of quality compared to the NR bench-mark. Because the quality of the new product directlyimpacts its remanufacturing cost, by increasing qualitythe OEM also increases the cost of remanufacturing.Thus, the IR cannot recover its cost due to its signifi-cant cost-to-value disadvantage and stays out of themarket. Table 1 shows that the OEM needs to increasequality to deter entry when the IR becomes a biggerthreat as a result of more favorable cost-to-value ratio.
Table 1 Equilibrium Product Quality and New and RemanufacturedProduct Quantities
Figure 2 graphically demonstrates how the OEM’schosen quality level depends on the IR’s cost ofremanufacturing. We refer to the OEM’s behavior inregion R2 as entry deterrence because the OEM pre-vents the IR’s entry by deviating from the NR bench-mark. Note that entry deterrence does not exist in theexogenous quality benchmark (see section A in theOnline Supplement). In other words, quantity alone isnot sufficient to deter entry.In region R3, the OEM can no longer deter the IR’s
entry. In this region, the OEM can choose a higher orlower quality level when compared to the NR bench-mark depending on who has the cost-to-value advan-tage. When the OEM has the cost-to-value advantage,that is, ad [ 1, it chooses a higher quality level. In thiscase, increasing product quality increases the reman-ufacturing cost but customer perception of remanu-factured product does not increase proportionally,which in turn weakens the IR’s competitive position.In contrast, when IR has the cost-to-value advantage,that is, a
d \ 1, the OEM chooses a lower quality levelto de-emphasize its competitor’s advantage.Finally, in region R4, the IR is very powerful and
remanufactures all available cores. In this region,there is little perceived quality difference between thenew and remanufactured products due to high d, andthe OEM cannot compete with the IR using the qual-ity lever. Thus, the OEM keeps the quality at the NRbenchmark level and instead competes with the IR bylimiting the new product quantity and thereby theavailable cores for remanufacturing. We call this thequantity-limiting strategy. Figure 2b shows theOEM’s quantity in the base model as well as in theNR and exogenous quality benchmarks. The figureillustrates that the OEM reduces the new productquantity compared to the NR benchmark to restrainthe IR. Because the OEM stops using the quality leverand instead focuses on the quantity lever in regionR4, there is discontinuity in the quantity and the qual-ity levels in Figure 2 when moving from region R3 toR4 due to this strategy switch. We do not observe the
same phenomenon in the order quantity of the exoge-nous quality benchmark shown in Figure 2b. This isbecause quantity is the only lever in the exogenousquality benchmark; therefore there is no switchingbetween strategies.Proposition 1 demonstrates when the IR has a sig-
nificant cost-to-value advantage, the OEM focusesprimarily on a quantity-limiting strategy. Indeed, thisis consistent what we see in the printer cartridgeindustry where IRs have a significant cost-to-valueadvantage and OEMs mostly try to compete with IRsby creating quantity scarcity.Proposition 1 illustrates that when the OEM has the
cost-to-value advantage, it relies more on the qualitylever, whereas when the advantage shifts to the IR, itincreasingly relies on the quantity lever. Regions R2and R4 demonstrate two extremes. In R2, the OEMhas a significant cost-to-value advantage, and it reliessolely on quality to deter the IR’s entry (the OEM’squantity is the monopoly quantity given its chosenquality). In contrast, in R4, the IR has a significantcost-to-value advantage, and the OEM uses only thequantity lever in this case, keeping its quality at theNR benchmark level.It is worthwhile to contrast our findings with the
monopoly remanufacturing benchmark. A monopo-list always increases its chosen product quality afterengaging in remanufacturing (details of the analysisare provided in section A.2 of the Online Supple-ment). In contrast, when remanufacturing is per-formed by the IR, the OEM can decrease productquality compared to the NR benchmark. This happensbecause the OEM’s quality decision directly affectsthe IR’s competitive position, while a monopolistOEM who remanufactures in-house does not need toworry about a competitor. When faced with competi-tion from an IR, the OEM needs to take into accountwho has the cost-to-value advantage when making itsquality decision. Under different modeling assump-tions, Atasu and Souza (2011) also find that a mono-polist OEM engaging in quality recovery (of which
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.3
0.32
0.34
0.36
0.38
0.4
0.42
α
Pro
duct
Qua
lity
R4 R3 R2 R1
No−reman.Base model
Quality
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10.27
0.28
0.29
0.3
0.31
0.32
0.33
0.34
α
New
Pro
duct
Qua
ntity
1R2R3R4R
No−reman.ExogenousBase model
New Product Quantity(a) (b)
Figure 2 Equilibrium New Product Quality and Quantity (d = 0.4, b = 1, Exogenous Quality sf ¼ 13b)
remanufacturing is an example) chooses a qualitylevel that is weakly higher than the no-recoveryscenario.
5. Consumer and Social Welfare
In this section we investigate the impact of remanu-facturing and quality choice on consumer surplus(CS) and social surplus (SS). The consumer surplus isgiven by
CS ¼Z1�qn
1�qn�qr
ðdhs� prÞdhþZ1
1�qn
ðhs� pnÞdh; ð1Þ
where the first term is the surplus from remanufac-tured products, and the second term is the surplusfrom new products sold. The social surplus is thesum of the consumer surplus and the firm profits.Intuition suggests that remanufacturing should
improve consumer welfare. Indeed, the next proposi-tion confirms this conjecture for the exogenous qual-ity benchmark, that is, when the OEM responds to theIR’s entry only with its quantity keeping its qualityconstant.
However, the OEM does not keep its product qual-ity constant when faced with the IR threat. Proposi-tion 1 shows how the OEM adjusts its product qualityto strengthen its competitive position. Basically, itmay choose lower or higher quality levels dependingon its cost-to-value position relative to the IR. Thenext proposition demonstrates that remanufacturingcan hurt CS due to the OEM’s quality choice.
PROPOSITION 3. There exists ac satisfying 1\ acd \ 8�d
4þdsuch that CS is higher than that of the NR benchmark ifand only if a\ ac. Furthermore, CS is strictly smallerthan that of the NR benchmark when ac
d \ ad \ 2.
Propositions 1 and 3 show acd falls in region R3. Thus,
CS is lower than or equal to the NR benchmark inregions R1 and R2. In region R1, the IR is not a threatand the outcome is identical to the NR benchmark. Inregion R2, however, CS is strictly smaller than that ofthe NR benchmark as shown in the second half of theproposition. Specifically, in region R2, the OEM ineffi-ciently chooses higher quality to deter the IR’s entryand therefore focuses on the higher valuation consum-ers, which in turn reduces CS. Interestingly, CS canalso suffer in region R3 even when the IR enters themarket. This is again due to the OEM’s choice of highquality to play to its cost-to-value advantage.
Proposition 3 indicates that an IR with a weak com-petitive position is not preferable for CS. In order forCS to benefit from remanufacturing, the IR must havea sufficiently strong cost-to-value advantage; other-wise the OEM’s quality response hurts CS. Thisdynamic does not exist and CS always increases withremanufacturing in the exogenous quality bench-mark. Thus, Propositions 2 and 3 imply that disre-garding the OEM’s quality decision can lead tooverestimating the benefit of remanufacturing forconsumers. Now let us consider the impact of reman-ufacturing on social surplus.
PROPOSITION 4. There exists as satisfying 1\ asd \ 8�d
4þdsuch that SS is higher than that of the NR benchmark ifand only if a\ as. Furthermore, SS is strictly smallerthan that of the NR benchmark when as
d \ ad \ 2.
The proposition indicates that the IR’s entry threatas well as its successful entry can decrease not onlyCS but also SS when the IR’s cost-to-value position isnot sufficiently favorable. Figure 3 compares SSagainst the NR and exogenous quality benchmarks. Inthe exogenous quality benchmark, the new productquality is kept at the NR benchmark quality disre-garding the OEM’s quality response to the IR threat.When the IR remanufactures, SS is always lower thanthe exogenous quality benchmark. Furthermore, notethat when 0.516 < a < 0.55, remanufacturing worsensSS in our base model while improving it in the exoge-nous quality benchmark. In this case, ignoring theOEM’s quality decision leads to incorrectly conclud-ing that remanufacturing would benefit social wel-fare.Propositions 3 and 4 show that an IR’s remanufac-
turing can decrease CS and SS. In contrast, ourmonopoly remanufacturing benchmark demonstratesthat both CS and SS increase when a monopolisticOEM engages in remanufacturing. Likewise, our
0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.80.0535
0.054
0.0545
0.055
0.0555
0.056
0.0565
0.057
R3 R2
α
Soc
ial S
urpl
us
No−reman.ExogenousBase model
Figure 3 Comparison of Social Surplus with No-Remanufacturing andExogenous Quality Benchmarks (d = 0.4, b = 1, and Exo-genous Quality sf ¼ 1
extension in section 7.1 (when only the OEM remanu-factures) show a similar result. This contrast is due totwo factors. First, an IR chooses to remanufactureproducts even when remanufacturing does not have avery attractive cost-to-value position. The OEMwould not choose to remanufacture in regions inwhich the IR’s remanufacturing decreases CS and SS.1
In other words, the OEM’s remanufacturing incen-tives are better aligned with consumer and social wel-fare compared to the IR’s. Second, when the OEMutilizes the benefits of remanufacturing, it choosesproduct quality more efficiently as far as CS and SSare concerned. In contrast, when an IR does theremanufacturing, the OEM can inefficiently increasequality to deter entry or decrease quality to under-mine the cost-to-value advantage of its competitor.Our findings have important policy implications.
There is an ongoing policy debate about whether andwhen to promote remanufacturing. For example, theRecycling/Remanufacturing Tax Credit Bill, HR 5695(The Remanufacturing Institute 2008) introduced bythe Automobile Parts Remanufacturers Association(APRA) calls for tax credits for investments in reman-ufacturing equipment. Although the bill did not passthe first time a round, efforts to pass it continue. Simi-larly, the Waste Electrical and Electronic Equipment(WEEE) Directive legislation in the European Unionholds manufacturers financially responsible for takingback and disposing of end-of-life electric and elec-tronic equipment. In a recent vote on changes to thedirective, a 5% reuse target was introduced to pro-mote higher levels of reuse/remanufacturing (Jowitt2011). In addition, environmental awareness cam-paigns, companies promoting sustainable businesspractices, etc. may work to improve customers’ per-ception of remanufactured products. Such incentivesand campaigns can alter competitive positioning ofIRs and OEMs and change their behavior. Our find-ings illustrate that policy makers should be carefulwhen designing such incentives especially when IRs(rather than OEMs themselves) engage in remanufac-turing. Making remanufacturing attractive for IRsdoes not necessarily improve social welfare. Proposi-tions 3 and 4 show that the IR’s threat and entry candecrease both CS and SS. Furthermore, ignoring com-petition or the OEM’s quality decision can lead tooverestimating benefits of remanufacturing for con-sumer and social surplus.
6. Environmental Impact
We follow the convention in the literature (Agrawalet al. 2012, Atasu and Souza 2011, White et al. 1999),and assume that one unit of new product and reman-ufactured product entail E and e environmentalimpact, respectively, considering all stages of product
life cycle, which includes production, use by custom-ers, end of life, and remanufacturing. Therefore, whenthe OEM produces qn units and the IR remanufacturesqr units, the total environmental impact is qnE þ qre.The next proposition shows the effect of remanufac-
turing on the environment comparing it to the NRbenchmark and describes how environmental impactdepends on relative cost a and perception d of theremanufactured product.
PROPOSITION 5. Table 2 shows how the environmentalimpact changes with a and d.
• When the IR is not a threat (region R1), the envi-ronmental impact is the same as the NR benchmarklevel.
• When the IR’s entry is deterred by the OEM(region R2), the environmental impact is alwayslower than the NR benchmark level.
• When the IR enters the market but does not reman-ufacture all available cores (region R3), the envi-ronmental impact is lower than the NR benchmarklevel if and only if e
E \ ð�2þaÞd2ð�8þdÞdþað4þdÞ.
• When the IR enters the market and remanufacturesall available cores (region R4), the environmentalimpact is lower than the NR benchmark level if andonly if e
E \ d2.
The proposition demonstrates that the IR’s entrythreat in region R2 reduces environmental impact. Todeter entry, the OEM increases product quality andfocuses on higher valuation customers, which in turndecreases the quantity sold. Furthermore, Table 2shows that as the IR becomes a bigger threat, the envi-ronmental impact decreases further in this regionsince the OEM needs to keep increasing quality todeter entry as the IR’s cost-to-value positionimproves.The IR remanufactures in regions R3 and R4,
and the relative impact of new and remanufacturedproducts e
E determines the environmental impact ofremanufacturing in these regions. Specifically, whenthe remanufactured product has a sufficiently smallerrelative environmental impact indicating small e
E, theoverall environmental impact decreases with the IR’sentry. Otherwise, remanufacturing increases the envi-ronmental impact.
Figure 4 illustrates how the environmental impactdepends on the IR’s relative competitive positionshowing that environmental impact attains its worstlevel in region R3. This is because competitionbetween the IR and the OEM is more intense, yieldingmore quantity sold (new + remanufactured) whenneither has a significant cost-to-value advantage. Theenvironmental impact gets smaller near region R2, asthe OEM’s cost-to-value advantage improves. Simi-larly at the other end, the environmental impact isalso smaller in region R4, where the IR has a signifi-cant cost-to-value advantage.2
In region R4, the OEM follows the quantity scarcitypolicy to limit the IR’s remanufacturing. Table 2shows that when the IR’s cost-to-value position getseven better due to a higher d in this region, the OEMfurther decreases its quantity, benefiting the environ-mental impact. In both regions R2 and R4, quantityand, hence, the environmental impact decrease whenthe IR becomes more powerful. But there are differentdynamics in place. In R2, quantity decreases becausethe OEM increases quality to deter the IR, whereas inregion R4, the OEM creates scarcity to limit the IR’sremanufacturing.Comparisons with the monopoly remanufacturing
benchmark (see section A.2 in the Online Supplementfor details) show that how remanufacturing changesthe environmental impact level depends on who—OEM or IR—does the remanufacturing. We find thatif remanufacturing has the cost-to-value advantage(ad \ 1) and, hence, is socially desirable,3 wheneverthe environmental impact in the base model is smallerthan the NR benchmark, environmental impact in themonopoly remanufacturing benchmark is also smal-ler than the NR benchmark but not vice versa. Hence,remanufacturing by an OEM is more likely todecrease environmental impact than remanufacturingby an IR (our extension in section 7.1 finds a similarresult). This is mainly due to two factors: (i) competi-tion increases the total quantity sold; a monopoly
always sells fewer units. (ii) The OEM can reduce thequality level when the competing IR has the cost-to-value advantage and a lower quality level implies abigger quantity in the market. Under somewhatdifferent modeling assumptions, Atasu and Souza(2011) find that quality recovery (of which remanufac-turing is an example) carried out by a monopolisticOEM always decreases the environmental impact,which is also in contrast with our base model. Ourfindings together with Atasu and Souza (2011) sug-gest that as far as the environmental impact is con-cerned, it may not be beneficial to encourage IRsrather than OEMs to remanufacture. Furthermore,when an IR does the remanufacturing, increased com-petition can aggravate environmental impact. In thiscase, it is desirable to have an IR with either a suffi-ciently unfavorable cost-to-value ratio so the OEMincreases the quality level or a sufficiently favorablecost-to-value ratio so the OEM competes by creatingquantity scarcity.
7. Additional Competitive Levers
In this section, we study two additional levers anOEM can use to compete with an IR. Specifically, theOEM can also remanufacture its own product or itcan collect cores to make them unavailable for the IR.
7.1. Remanufacturing by Both OEM and IRRemanufacturing can be done by IRs as well as by theOEM itself. There are examples of both in practice.For example, Xerox leases its copiers and remanufac-tures end-of-lease copiers by itself; in contrast, in thecartridge industry mainly IRs do the remanufactur-ing. Here, we extend our base model and allow theOEM to remanufacture its own product in addition tothe IR. We conduct a numerical study to analyze theresulting equilibrium.The OEM and the IR have the same remanufactur-
ing cost (bas2 in our model), and they choose their
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10.85
0.9
0.95
1
1.05
1.1
α
Impa
ct
2R3R4R
No−reman.Base model
Effect of α
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10.85
0.9
0.95
1
1.05
1.1
δ
Impa
ct
R1 R2 R3 R4
No−reman.Base model
Effect of δ(a) (b)
Figure 4 Environmental Impact (e = 1, E = 3, d = 0.5 in (a), a = 0.5 in (b))
desired remanufacturing quantities simultaneously.However, the OEM has the priority in quantity alloca-tion when their total demand exceeds the number ofavailable cores. In other words, the IR can remanufac-ture only the cores that the OEM chooses not toremanufacture. Admittedly, this approach favorsOEM’s remanufacturing, but even with this bias, weshow the OEM may prefer letting the IR remanufac-ture and instead continue to compete through manip-ulating quality. Note the other extreme where the IRgets priority in the allocation of available cores resultsin the same equilibrium outcome as our base model.4
Table 3 reports results of our numerical study as d var-ies for one a value, a = 0.8. In our study, we repeat thesame analysis for a ∈ {0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9}values and find that Table 3 is representative of theiroutcomes as well. Quality cost coefficient b is a scale fac-tor in our model and it is kept at b = 1. In the table, qmrand qir show the number of remanufactured units bythe OEM and the IR, respectively. Furthermore, e/Eshows the maximum e/E ratio—environmental impactof remanufactured product relative the new product–below which remanufacturing (or the possibility of it)reduces environmental impact compared to the NRbenchmark. In the table e/E is not reported for d = 0.4since when d � 0.4, remanufacturing is not viable, andthe NR benchmark and our extended model yields thesame outcome.To better understand the effect of competition, con-
sider the OEM’s optimal policy in the absence of anIR. A monopolist OEM does not remanufacture whenthe cost-to-value ratio favors the new product, that is,a/d > 1. It remanufactures some but not all availablecores when the remanufactured product has the cost-to-value advantage, that is, a/d < 1 but the advantageis not sufficiently big (0.8 < d < 0.9 in Table 3).Finally, when the remanufactured product has a sig-nificant cost-to-value advantage, a monopolist OEMremanufactures all available cores. Table 3 shows thatwhen remanufacturing has a sufficiently big advan-tage or disadvantage, the OEM does not need to devi-ate from the monopoly optimal policy to competewith the IR. Specifically, when remanufacturing has asevere disadvantage (d � 0.4), the IR is not a threatand the OEM sells only the new product. In contrast,when remanufacturing has a significant advantage(d � 0.91), the OEM remanufactures all availablecores, leaving no cores to the IR.When cost-to-value ratio of remanufacturing a/d is
moderate (0.4 < d � 0.88 in Table 3), the OEM needsto actively compete with the IR. The OEM usesdifferent policies depending on the cost-to-valueposition of remanufacturing. Note that remanufac-turing becomes increasingly attractive as d increases.When 0.40 < d < 0.49, similar to our base model, theOEM increases quality to deter the IR from
remanufacturing. When 0.49 � d < 0.61, the OEMlets the IR remanufacture but it increases quality toweaken the IR’s competitive position. It is interestingthat the OEM is using only quality as a strategiclever in 0.40 < d < 0.61, although our core allocationgives absolute priority to the OEM. Finally, the OEMinefficiently remanufactures all available cores itselfin order to leave no cores available to the IR when0.61 � d � 0.81. This discussion demonstrates thatsimilar to our base model, the OEM relies on qualityas a competitive lever when remanufacturing doesnot have a strong cost-to-value position; in contrast,it uses a quantity limiting strategy when the IR’sremanufacturing becomes a bigger threat.The OEM increases quality when the remanufac-
tured product has the cost-to-value advantage, that is,a/d < 1. This is in direct contrast to the base model.Essentially, when the OEM itself rather than a com-petitor IR does the remanufacturing, the OEM isbetter off underscoring the remanufactured product’sadvantage by increasing quality. However, when theremanufactured product has the disadvantage, thatis, a/d > 1, and the OEM remanufactures solely toeliminate available cores for the IR, the OEMdecreases quality.Similar to the base model, CS and SS decrease when
the OEM uses quality to deter the IR’s entry(0.40 < d < 0.49).The NR benchmark is equivalent tothe d = 0.4 outcome. Likewise, the IR’s remanufactur-ing can also decrease CS and SS (d = 0.49). In theseexamples, the OEM inefficiently chooses a high qual-ity level to strengthen its competitive position. Simi-larly, CS and SS suffer when 0.61 � d � 0.85 and
Table 3 Equilibrium and the Resulting Consumer/Social Surplus andEnvironmental Impact when the OEM Can Also Remanufacture(b = 1, a = 0.8)
the OEM inefficiently remanufactures all availablecores itself to starve the IR in this range. When thecost-to-value position of remanufacturing improves(d � 0.88), the OEM’s remanufacturing increasesboth CS and SS compared to the NR benchmark.When the OEM does not remanufacture, the envi-
ronmental impact is the same as our base model, andour insights carry over. However, contrasting theenvironmental impact of OEM’s and IR’s remanufac-turing generates an additional insight. Remanufactur-ing decreases the environmental impact when e/E issmaller than e/E in Table 3. Thus a larger e/E indi-cates that remanufacturing is more likely to reducethe environmental impact. Improving the cost-to-value ratio of remanufacturing (higher d in the table)decreases e/E when the IR is remanufacturing andincreases e/Ewhen the OEM is remanufacturing. Thissuggests that making remanufacturing more attrac-tive can worsen the environmental impact whenremanufacturing is done by the IR whereas it lessensthe environmental impact when remanufacturing isdone by the OEM itself.
7.2. Preemptive CollectionIn our base model, the OEM competes with the IRusing quality and quantity as strategic levers. Here, inaddition to using quality and quantity, we allow theOEM to collect and dispose of cores to compete withthe IR. As before, the OEM first chooses the qualitylevel. Then simultaneously, the OEM decides thenumber of cores to collect for disposal and the newproduct quantity and the IR decides the remanufac-tured product quantity. The OEM has priority in corecollection (i.e., it has first access to cores) if the totaldemand for cores exceeds the available cores. Eventhen, we show that the OEM may still rely on qualityto compete with the IR rather than collecting and dis-posing of cores. Similar to Ferguson and Toktay(2006), we assume that the total collection and dis-posal cost the OEM incurs is hq2d where qd is the quan-tity collected and h is a measure of how difficult andexpensive it is to collect cores. Due to the analyticalcomplexity of this model, we conduct a numericalstudy.Figure 5 illustrates the OEM’s quality choice and equi-
librium regions for d = 0.4 and h = 0.04 when a variesfrom 0 to 1. In region Rd the OEM collects all availablecores and the regions R1�R3, in which the OEM doesnot utilize preemptive collection, are the same as those ofour base model. We repeat the numerical study for allcombinations of d = {0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9} andh = {0.04,0.05,0.06,0.07,0.08,0.09,0.10,0.11} and observethat the figure is a representative outcome.The figure shows that when the cost-to-value ratio a
dis sufficiently high (0.59 � a < 0.8), the OEM usesquality to compete with the IR instead of preemptive
collection (The IR does not pose a threat whena � 0.8). When 0.69 � a < 0.8, the OEM detersthe IR’s entry by increasing quality. When0.59 � a < 0.69 the OEM lets the IR remanufacturebut still chooses a high quality level to weaken the IR.Drivers of these results are the same as those in thebase model. When the cost-to-value ratio a
d is suffi-ciently small (0 < a < 0.59), the IR’s competitive posi-tion is strong. In this case, the OEM collects anddisposes of all available cores to deter the IR’s entry.While doing so, the OEM also increases quality rela-tive to the NR benchmark to decrease the number ofcores to be collected. Hence, the OEM utilizes the pre-emptive collection and quality levers together to deterIR’s entry.When the OEM uses quality to deter or compete
with the IR (i.e., 0.59 � a < 0.8), the threat or actualentry can decrease the CS and SS compared to the NRbenchmark. This behavior is similar to our basemodel. For 0 < a < 0.59, the OEM uses preemptivecollection to deter the IR’s entry, and CS and SS arelower than the NR benchmark levels. This behavior isalso consistent with our base model where entrydeterrence reduces CS and SS levels. In section B.1 ofthe Online Supplement, we provide further details onour social welfare results.In the numerical study we observe that when h is
high (h � 0.09), collecting all available cores may notbe viable. In this case, the OEM collects and disposesof a fraction of the available cores and the IR remanu-factures the remaining cores. On the other hand,when h is very low, as intuition would suggest, theOEM collects and disposes of all cores.
7.3. Comparison of Competitive LeversThrough a numerical study, we now discuss how theOEM chooses to compete with the IR when all threecompetitive levers, that is, quality choice, remanufac-turing in-house, and preemptive collection, are avail-able. In our study, we considered all combinations of
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.32
0.33
0.34
0.35
0.36
0.37
0.38
0.39
0.4
α
Qua
lity
Rd R3 R2 R1
No−reman.h=0.04
Figure 5 Equilibrium Quality when the OEM Can Collect and Disposeof the Used Cores (d = 0.4, b = 1, h = 0.04)
a ∈ {0.1,0.2,…,0.8,0.9} and h ∈ {0.04,0.05,0.06}. Table4 is representative of our results.Similar to our earlier results, the OEM’s choice
depends on the remanufactured product’s relativecost-to-value ratio a
d. Consistent with our insights fromthe base model, when the cost-to-value ratio is highbut the remanufactured product is still a competitivethreat, the OEM relies only on the quality lever tocompete with the IR. Specifically, when d = 0.5, theOEM allows the IR to remanufacture but increasesproduct quality relative to the NR benchmark toundermine the IR’s competitive position. Likewise,when d = 0.45, the OEM increases quality relative tothe NR benchmark to deter the IR’s entry. The IR isnot a competitive threat when d � 0.4.When the remanufactured products’s relative cost-
to-value ratio is low, that is, the IR becomes a biggercompetitive threat, the OEM uses in-house remanu-facturing and preemptive collection jointly to causescarcity of cores. In particular, when d � 0.6, theOEM remanufactures a fraction of the available coresand preemptively collects any remaining cores, deter-ring the IR’s entry. Furthermore, the OEM remanufac-tures a larger proportion of collected cores when dincreases, indicating a higher perceived value for theremanufactured product. This result is in agreementwith our insight from the base model, in which theOEM decreases the production of new product tolimit the available cores when the IR becomes a biggerthreat.
8. Extensions
8.1. Price CompetitionHere, we study what happens when the OEM and theIR compete in prices. The following propositiondescribes the equilibrium for the price competitiongame showing that the structure of the equilibrium isthe same as the quantity game.
PROPOSITION 6. The following characterizes the equilib-rium regions when the OEM and the IR compete inprices.
R1p: If ad � 2, the IR cannot enter the market and the
OEM acts like a monopoly.
R2p: If 4�d2þd � a
d \ 2, the IR is a threat and its entryis deterred by the OEM.
R3p: If dð10�dÞð4�dÞ2 \ a
d \4�d2þd, the IR enters but does not
remanufacture all available cores.
R4p: If 0\ ad � dð10�dÞ
ð4�dÞ2 , the IR enters the market andremanufactures all available cores.
The equilibrium quality, new and remanufactured prod-uct prices, and quantities are provided in the proof of theproposition.
Regions R1p–R4p are the same as regions R1–R4 ofour base model. Specifically, in region R1p, the IR isnot a threat due to its poor cost-to-value position. Inregion R2p, the OEM chooses a higher quality levelcompared to the NR benchmark to deter the IR’sentry. In region R3p, the OEM chooses a higher orlower quality level depending on whether it has thecost-to-value advantage or disadvantage. Finally, inregion R4p, the OEM follows a quantity limiting strat-egy. Drivers of these results are the same as those inour base model. In region R2p, the OEM’s price issmaller than the monopoly price for its chosen prod-uct quality. Different from our base model, the OEMuses price in addition to quality to deter entry inregion R2p.It is well known that price competition is more
intense than quantity competition and leads to higherCS and SS (Singh and Vives 1994). Consistent withthis fact, we find that CS and SS are higher than theNR benchmark when the OEM and the IR compete inprices (more detailed analysis of the CS and SS underprice competition is relegated to section B.2 of theOnline Supplement). Another artifact of the intensecompetition is that the new product quantity isalways higher than or equal to the NR benchmark.Therefore, remanufacturing by an IR always increasesenvironmental impact under price competition.
8.2. Alternative Remanufacturing CostUp to this point, we assumed that all remanufacturingrelated costs are subsumed in bas2. In this section, weconsider an additional cost term n that is independentof the quality level. Specifically, the IR’s total unitremanufacturing cost becomes bas2 þ n.We are able to characterize the equilibrium when the
OEM has the cost-to-value advantage, that is, ad � 1, andwe state our result in Proposition 7. However, when theIR has the cost-to-value advantage, that is, a
d \ 1, themodel is not analytically tractable; therefore we resort toa numerical study. Figure 6 demonstrates the results forad \ 1 as well as for a
d � 1. While the figure reports theresult for one d and n = {0,0.01,0.02,0.05,0.06}, we haverun the numerical study for all combinations ofd ∈ {0.2,0.3,0.4,0.5,0.6,0.7,0.9} and n ∈ {0.005,0.010,0.020,0.025,0.030,0.035,0.040,0.045,0.050} and have found that
Table 4 Equilibrium when the OEM Can Remanufacture andPreemptively Collect (b = 1, a = 0.8, h = 0.04)
they are all consistent. We also study the impact of thequality independent remanufacturing cost on the CSand SS in the Online supplement (see section B.3) andobserve numerically that the insights from Propositions3 and 4 continue to hold.
PROPOSITION 7. The following characterizes the equilib-rium regions for a
d � 1 when the IR incurs an additionalcost n per unit.
R1i. If ad � 2 or 2 [ a
d [ 1 and n � 2d�a9b , the IR
cannot enter and the OEM acts like a monopoly.
R2ai. If 2 [ ad � 8�d
4þd and 2d�a9b [ n or 8�d
4þd [ ad � 5
4and 2d�a
9b [ n � n0, the IR’s entry is deterred by theOEM, who chooses a quality level higher than the NRbenchmark.
R2bi. If 54 [ a
d � 1 and 2d�a9b [ n � n0, the IR’s
entry is deterred by the OEM who chooses a qualitylevel lower than the NR benchmark.
R3i. If 8�d4þd [ a
d � 1 and n0 [ n, the IR enters themarket but does not remanufacture all available cores.
The equilibrium quality, new and remanufactured prod-uct quantities, and n0 are stated in the proof of the Prop-osition.
Regions R1i–R3i are same as the regions R1–R3 inthe base model. The proposition demonstrates thatall three regions that exist in our base model forad � 1, namely R1�R3, continue to exist. In additionto these regions, an additional region (region R2bi)where the OEM deters the IR’s entry by choosing aquality level lower than the NR benchmark is alsopossible when the cost-to-value ratio and the qualityindependent remanufacturing cost are at moderatelevels, that is, �aþ2d
9b [ n � n0. The OEM’s choice oflow quality decreases the demand for the remanu-factured product but also decreases the remanufac-turing cost. The key point is that the qualityindependent component (n) of the remanufacturing
cost does not change when the OEM chooses a lowquality level, and, therefore, the positive effect ofcost reduction on the IR’s profit is smaller whencompared to the negative effect of demand reduc-tion. This allows the OEM to deter the IR’s entrythrough decreasing quality in the presence of thequality independent cost component. The proposi-tion also demonstrates that when the quality inde-pendent remanufacturing cost is too high, that is,2d�a9b � n, the IR cannot enter at all, as expected.Figures 6a and 6b illustrate the equilibrium struc-
ture for n ∈ {0,0.01,0.02} and n ∈ {0.05,0.06} respec-tively. Figure 6a shows that when the IR has astrong cost-to-value position, the OEM may continueto rely on reducing production and limiting coreavailability (region R4i). However, as intuition sug-gests, region R4i gets smaller as n increases. In fact,when n � 0.02, R4i disappears. Figure 7 also showsthat as n increases, the OEM relies more on thequality lever to compete with the IR. However as nincreases, the regions where the OEM chooses aquality level higher than the NR benchmark shrink.In fact, for n � 0.05, the OEM always chooses alower level of quality (if different from the NRbenchmark level).
8.3. Independent Quality GapIn our base model the quality gap between the newand remanufactured product is proportional to theproduct quality s. Here, we consider an alternativemodel in which the quality gap is independent ofproduct quality; specifically, the value of the remanu-factured product is h(s�/) for type-h consumer,where / shows the quality gap for the remanufac-tured product.Due to the analytical complexity of this alterna-
tive model, we resort to numerical studies. Figure 7shows the equilibrium quality and quantity as qual-ity gap / varies for a = 0.4. We find the behavior in
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.24
0.26
0.28
0.3
0.32
0.34
0.36
0.38
0.4
0.42
α
Qua
lity
R4(n=0)
R3(n=0)
R2a(n=0)
R1(n=0)
n=0n=0.01n=0.02
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.22
0.24
0.26
0.28
0.3
0.32
0.34
α
Qua
lity
R3(n=0.05)
R2b(n=0.05)
R1(n=0.05)
n=0.05n=0.06
(a) (b)
Figure 6 Equilibrium Quality Level when the IR Incurs Quality-Independent Cost (d = 0.4, b = 1); (a) n ∈ {0,0.01,0.02} (Partitions R1–R4 Are Shownfor n = 0); (b) n ∈ {0,0.05,0.06} (Partitions R1–R3 Are Shown for n = 0)
this figure to be robust by also checking othera ∈ {0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9} values. The figureidentifies four regions similar to our base model(see Proposition 1). In particular, in region R1, thequality gap is sufficiently high and the IR is not athreat. In region R2, the OEM deters the IR’s entrythrough its quality choice. In region R3, the qualitygap is sufficiently small and the IR remanufacturesa portion of available cores. In region R4, the qual-ity gap is very small, and the OEM follows a quan-tity-limiting strategy. This strategy shift is evidentin Figure 7b, as the quantity drops discontinuouslybetween regions R3 and R4. Note that similar to ourbase model, when the IR is weak (large / in thisextension), the OEM competes using the qualitylever; in contrast, when the IR is strong (small /),the OEM relies on limiting quantity.Figure 7a demonstrates that the OEM always
chooses a lower quality level compared to the NRbenchmark. This is the main difference between thisextension and our base model. Because the qualitygap is independent of the quality level, increasing thequality of the new product also increases the qualityof the remanufactured product by the same amount.Therefore, the OEM does not want to increase qual-ity too much, which would undermine the relativesignificance of the quality gap. A lower quality levelensures that the OEM’s quality advantage is suffi-ciently large relative to the remanufactured product’sperceived quality. When the OEM chooses a muchlower quality level than the NR benchmark, this nega-tively affects social welfare and results in CS and SSlevels lower than the NR benchmark (a more detailedanalysis is provided in section B.4 of the OnlineSupplement).
9. Concluding Remarks
We study how an OEM can use product quality as acompetitive strategic lever along with quantityagainst an IR. Even though there is evidence that
OEMs take competition and remanufacturing intoconsideration in their product design decisions, thisproblem has not been studied before. The relationshipbetween quality and competition has been studied inthe economics and marketing literatures, but theirresults do not directly apply because the remanufac-turing context is fundamentally different. By charac-terizing how the OEM competes with the IR inequilibrium, we find that the OEM relies more onquality as a strategic lever when it has a stronger com-petitive position, and, in contrast, it relies more heav-ily on limiting quantity of cores when it has a weakercompetitive position.A commonly held belief is that remanufacturing is
good for the environment and consumers eventhough these relationships are not well understood,especially in industries where predominantly IRsremanufacture. We study the effect of remanufactur-ing by an IR on total environmental impact and con-sumer surplus. We find that unless the IR has asufficiently weak competitive position (so the OEMcan deter entry) or a sufficiently strong one (so theOEM switches its competitive strategy and limitsproduct quantity), environmental impact can increasewhen compared to the NR benchmark, because whenneither the OEM nor the IR has a sufficiently strongcompetitive advantage, the competition between thetwo becomes more intense, yielding more quantitysold (new + remanufactured). On the consumer sur-plus side, not only the IR’s entry threat but also itssuccessful entry can cause a decrease in the consumersurplus level. This is also in contrast with our monop-oly remanufacturing benchmark, which shows thatremanufacturing by an OEM always benefits the con-sumers (Atasu and Souza 2011 have a similar find-ing). Taken together, our findings regardingenvironmental impact and social welfare suggest thatpolicy makers should be careful about promoting IRremanufacturing over OEM remanufacturing.Some limitations of our work are worth mention-
ing. We study a single period model due to its
0.05 0.1 0.15 0.2 0.25 0.3
0.18
0.2
0.22
0.24
0.26
0.28
0.3
0.32
φ
Qua
lity
1R2R3R4R
No−reman.Base model
Quality
0.05 0.1 0.15 0.2 0.25 0.3
0.24
0.26
0.28
0.3
0.32
0.34
0.36
0.38
0.4
φ
New
Pro
duct
Qua
ntity
1R2R3R4R
No−reman.Base model
New Product Quantity(a) (b)
Figure 7 Equilibrium Quality and New Product Quantity when the Perceptual Quality Gap Is Independent of New Product Quality (a = 0.4, b = 1)
tractability and to keep our focus on our researchquestions. This approach is plausible when a prod-uct’s pay-off during its mature stage makes up a bulkof its total payoffs. Indeed, most studies looking atfirms’ quality choice consider stationary demand aswe do (e.g., Atasu and Souza 2011, Johnson and Myatt2006, Netessine and Taylor 2006, Plambeck and Wang2009). The relation between the shape of a product’slife cycle and its remanufacturing decisions can be aninteresting research question, which we do notaddress in this study and leave for future work. Fur-thermore, comparison of our results from the basemodel with results from extensions where weconsider alternative cost and consumer valuationfunctions for the remanufactured product indicatethat whether the OEM chooses to increase or decreasethe quality level vis-à-vis the NR benchmark can besensitive to the functional form assumed. Similarly,the implication of remanufacturing on the social andconsumer surplus, and environmental impact can besensitive to the form of competition (price vs. quan-tity). Future research on this issue should be carefulabout these relations.
Acknowledgments
The authors thank Gilvan Souza, the senior editor and thereferees for many helpful comments and suggestions.
Notes
1The monopoly remanufacturing benchmark in the OnlineSupplement illustrates that the OEM never remanufactureswhen the remanufactured product has an inferior cost-to-value position compared to the new product.2Although e < E in Figure 4, the insights discussed herehold for e � E as well.3We know from section 5 and section A.2 in the Online Sup-plement that in both the base model and the monopolyremanufacturing benchmark, CS and SS levels are higherthan the NR benchmark when a
d \ 1. In addition, in themonopoly remanufacturing benchmark, the OEM remanu-factures only when it is socially advantageous to do so.4Essentially, in this scenario, any core that is not profitablefor the IR to remanufacture is not profitable for the OEMeither. Therefore, in equilibrium remanufacturing is doneonly by the IR, which is the same as our base model.However, when the OEM has the priority, a core that isnot profitable for the OEM can be profitable to remanufac-ture for the IR since, unlike the OEM, the IR does notneed to worry about cannibalization of the new product.5The NR benchmark is equivalent to the d = 0.4 outcome.
ReferencesAgrawal, V., A. Atasu, K. Van Ittersum. 2011. The effect of reman-
ufacturing on the perceived value of new products. Workingpaper, Georgia Institute of Technology, Atlanta, GA.
Agrawal, V., M. Ferguson, L. B. Toktay, V. Thomas. 2012. Is leas-ing greener than selling? Manage. Sci. 58(3): 523–533.
Atasu, A., V. D. Guide, L. VanWassenhove 2010. So what ifremanufacturing cannibalizes new product sales? Calif.Manag. Rev. 52(2): 56–76.
Atasu, A., M. Sarvary, L. N. Van Wassenhove. 2008. Remanufac-turing as a marketing strategy. Manag. Sci. 54(10): 1731–1746.
Atasu, A., G. D. Souza. 2011. How does product recovery affectquality choice? Prod. Oper. Manag. Advance online publica-tion. doi: 10.1111/j.1937-5956.2011.01290.x.
Atasu, A., L. N. Wassenhove, M. Sarvary. 2009. Efficient take-backlegislation. Prod. Oper. Manag. 18(3): 243–258.
Banker, R. D., I. Khosla, K. K. Sinha. 1998. Quality and competi-tion. Manage. Sci. 44(9): 1179–1192.
Bernstein, F., A. Federgruen. 2004. A general equilibrium modelfor industries with price and service competition. Oper. Res.52(6): 868–886.
Debo, L. G., L.B. Toktay, L. N. VanWassenhove. 2005. Market seg-mentation and technology selection for remanufacturableproducts. Manage. Sci. 51(8): 1193–1205.
Desai, P. S. 2001. Quality segmentation in spatial markets:Whendoes cannibalization affect product line design. Market. Sci.20(3): 265–283.
Elsberry, R. B. 2002. New York City Electronic Equipment Recyclingand Reuse act. Available at http://www.imagingeconomics.com/issues/articles/2002-09_10.asp (accessed date November20, 2011)
Esenduran, G., E. Kemahlioglu-Ziya, J. M. Swaminathan. 2010.The impact of takeback legislations on remanufacturing.Working paper, Kenan-Flagler Business School, University ofNorth Carolina, Chapel Hill, NC.
Esenduran, G., E. Kemahlioglu-Ziya, J. M. Swaminathan. 2012.Increased remanufacturing through take-back legislation andcompetition. Working Paper, Kenan-Flagler Business School,University of North Carolina, Chapel Hill, NC.
Ferguson, M., L. B. Toktay. 2006. The effect of competition onrecovery strategies. Prod. Oper. Manag. 15(3): 351–368.
Ferrer, G., J. Swaminathan. 2006. Managing new and remanufac-tured products. Manage. Sci. 52(1): 15–26.
Ferrer, G., J. Swaminathan. 2010. Managing new and differenti-ated remanufactured products. Eur. J. Oper. Res. 203(2):370–379.
Fishman, A., R. Rob. 2000. Product innovation by a durable-goodmonopoly. Rand J. Econ. 31(2): 237–252.
Gutowski, T. G., S. Sahni, A. Boustani, S. C. Graves. 2011. Reman-ufacturing and energy savings. Environ. Sci. Tech. 45(10):4540–4547.
Hauser, W., R. Lund. 2008. Remanufacturing: Operating practicesand strategies. Technical report, Boston University, Boston,MA.
Heese, H. S., K. Cattani, G. Ferrer, W. Gilland, A. V. Roth. 2005.Competitive advantage through take-back of used products.Euro. J. Oper. Res. 164(1): 143–157.
HP. 2010a. HP Planet Partners Recycling Program. Availableat http://www.hp.com/hpinfo/globalcitizenship/environment/recycling/product-recycling.html (accessed date February 23,2011).
HP. 2010b. HP Resolves Ink Cartridge Patent Infringement Com-plaint. Available at http://www.hp. com/hpinfo/newsroom/press/2010/100318a.html (accessed date February 23, 2011).
Johnson, J. P., D. P. Myatt. 2006. Multiproduct Cournot oligopoly.Rand J. Econ. 37(3): 583–601.
Jowitt, T. 2011. Europe votes for tougher WEEE recycling rules.eWeek Europe Feb. 7. Available at http://www.techweekeurope.co.uk/news/europe-votes-for-tougher-weee-recycling-rules-20068 (accessed date October 30, 2012)
Lexmark. 2010. Lexmarks Killer Chip Kills Reuse. Available at http://www.grrn.org/lexmark/ background.html. (accessed date Febru-ary 23, 2011).
Majumder, P., H. Groenevelt. 2001. Competition in remanufactur-ing. Prod. Oper. Manag. 10(2): 125–141.
Moorthy, K. S. 1988. Product and price competition in a duopoly.Market. Sci. 7(2): 141–168.
Motor and Equipment Remanufacturers Association. 2011. Manufac-tured Again—An Education and Awareness Campaign. Avail-able at http://www.mera.org/Advocacy/Manufactured-Again(accessed date November 23, 2011).
Netessine, S, T. Taylor. 2006. Product line design and productiontechnology. Market. Sci. 26(1): 101–117.
Plambeck, E., Q. Wang. 2009. Effects of e-waste regulation on newproduct introduction. Manage. Sci. 55(3): 333–347.
Robotis, A., T. Boyaci, V. Verter. 2009. Investing in product reus-ability: The effect of remanufacturing cost and demand uncer-tainities. Working paper, Desautels Faculty of Management,Montreal, Canada.
Singh, N., X. Vives. 1994. Price and quantity competition in a dif-ferentiated duopoly. Rand J. Econ. 15(4): 546–554.
Souza, G., B. Bayus, H. Wagner. 2004. New product strategy andindustry clockspeed. Manage. Sci. 50(4): 537–544.
Subramanian, R., M. Ferguson, B. Toktay. 2011. Component com-monality strategies under remanufacturing. Working paper,Georgia Institute of Technology – College of Management,Atlanta, GA.
The Remanufacturing Institute. 2008. Government Relations. Availableat http://www.reman.org/GovRelations_main.htm (accessed dateOctober 30, 2012).
Tsay, A., N. Agrawal. 2000. Channel dynamics under price andservice competition. Manuf. Serv. Oper. Manag. 2(4): 372–391.
US Environmental Protection Agency. 1997. Remanufacturedproducts: Good as new. WasteWi$e Update. Solid Waste andEmergency Response (5306W): EPA530-N-97-002.
White, A. L., M. Stoughton, L. Feng. 1999. Servicizing: The quiettransition to extended producer responsibility. Technicalreport. U.S. Environmental Protection Agency Office of SolidWaste. Available at http://www.greenbiz.com/sites/default/files/document/O16F7332.pdf (accessed date October 30,2012).
Supporting InformationAdditional Supporting Information may be found in theonline version of this article:
Appendix A: Benchmarks
Appendix B: Consumer and Social Welfare Results forExtensions to the Base Model
