1-s2.0-S0925527315001681-main

Pricing strategy of environmental sustainable supply chainwith internalizing externalitiesHuiping Dingn, Qilan Zhao, Zhirong An, Jia Xu, Qian LiuSchool of Economics and Management, Beijing Jiaotong University, Beijing 100044, Chinaa rti cle in foArticle history:Received 15 April 2014Accepted 7 May 2015Keywords:Sustainable supply chainExternalityEnvironmental performancePrice strategyGovernment policyabstractThe negative impact of business activities on resources and sustainability include pollution andenvironmental externalities, which are becoming more severe and attracting worldwide attention. Thisstudy investigates a mechanismfor motivating supplychainmembers tocollaboratively produceenvironmentallyfriendlyproducts(EFPs), withaninvestmentinpollutionreductionandpreventionandbyconsideringenvironmental regulationsandincentivepoliciesFromtheperspectiveof supplychain members, we focus on studying optimal pricing strategies for environmentally sustainable supplychains and the relationship betweenrm performance and environmental policy incentives. By treatingthe EFP as an investment project, its economic feasibility can be evaluated in the context of the supplychain. Unliketraditional supply chainmodels, inourmodel, governmentpolicyincentivesare sharedwithin the supply chain through transfer price negotiations between manufacturers and suppliers. Thisstudy addresses the impact of government policy incentives on value transition and prot allotment inthe collaborative supply chain system.& 2015 Elsevier B.V. All rights reserved.1. IntroductionWithincreasingenvironmental problems andthefailureoftreatment after pollution strategies, business activities haveaccumulated signicant negative externalities. These negativeenvironmental externalities refer to the damage caused by envir-onmental pollutiontothepublicwhenbusiness rmshavenotundertaken their responsibilities to compensate for the damage tothe society and environment. In traditional cost accounting,product cost only includes direct material cost, labor cost, manu-facturing cost, and so on; however, it does not consider theenvironmental resource cost to therm. Ignorance regarding hugeenvironmental externalitycost incurredbythe rms' environ-mentalunfriendlyactivitiesinthetraditionalaccountingsystemleads to market failure. Therefore, the mechanism and measures tocontrolandprevent environmentalpollutionshouldbeexploredat source. The internalization of negative externalities is the mosteffective method to correct market failure (Coase, 1960). In otherwords, rms are required to focus on preventing pollution at thesource, undertaking investment in environmental protection,decreasing or even eliminating pollution and environment damagecausedbytheproductionandconsumptionof therm'senvir-onmentally unfriendly products, and reducing or clearing the costsborne by the public. Environmental management is more than justseparate actions of an individual rm and requires the participa-tion and cooperation of groups ofrms. A focus on environmentalresponsibility in the supply chain leads to new ways of collaborat-ing with suppliers in terms of developing environmentally friendlytechnologies with critical suppliers, and providing training for allsuppliers from second-tier suppliers to sub-tier suppliers (Kovacs,2008). Assuming environmental responsibility in the context of asupply chainrequiresthejointeffortsofsupply chainmembers,government, and consumers and depends on the overall effective-ness and efciency of the supply chain members' collaborations. Itis meaningful tostudythecoordinationmechanismof havingbusiness rms participateinenvironmental technologyinvest-ment in the supply chain context. This reects pressure from thestakeholders' environmental interests and also inuences thevalue transmitted among the channel members.Internalizingenvironmentalexternality costrequires thesup-ply chainmembers to invest collaboratively inenvironmentalpollution prevention;consequently, thisleadsto additionalcostsand higher prices for environmental friendly products (EFPs)compared with environmentally unfriendly products (EUFPs),andpresentsachallengeforproducingEFPsintermsofcostorpricedisadvantageinthecompetitivemarket. Under situationswhere government regulations and incentive policies are lacking,supply chain rmsdriven by the pursuit of protlack theimpetus to invest in environmentally cleaner production. Asoverseer and supervisor of natural resources and the environment,Contents lists available at ScienceDirectjournalhomepage:www.elsevier.com/locate/ijpeInt. J. Production Economicshttp://dx.doi.org/10.1016/j.ijpe.2015.05.0160925-5273/& 2015 Elsevier B.V. All rights reserved.nCorresponding author. Tel.: 86 10 51687177; fax: 86 10 51683577.E-mail address: [email protected] (H. Ding).Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iInt. J. Production Economics () the governmenthasa responsibility to regulaterm activities infavor of environmental sustainability. Besides, imposing a penaltyon pollution activities, government should also induce and stimu-late rms toinvest inenvironmental protectionviaincentive-basedpolicies. Thenescollectedforpollutionactivitiescanbeused to stimulate environmental protection investment. Obviously,there are several questions that should be addressed in thisdomain. Forexample, whatare themajorfactorsandconditionsto have supply chain members participate in environmentalprotection investment? What are the effects of governmentregulations andpolicyincentives onthesupplychain's pricingdecisionsintermsof negotiatingtransfer pricebetweenmem-bers?Nevertheless, supplychainpricingisthecoreissuewhenpollutionpreventioninvestmentsarejointlymadebythemem-bers, andtheimpact of government policies onenvironmentalperformanceof thesupplychainisworthyof exploration. Thisstudy primarily focuses on supply chain pricing strategy from theperspective of motivating supply chain members cooperatively toproduceEFPsandtherebypromoteenvironmentalsustainability.As driving factors of promoting environmental technology invest-ment, regulations and policy incentives are explicitly introduced inthe context of supply chain management to study their impacts onthe supply chain's environmental performance. These issues drawpublicattention worldwide, particularlyindevelopingcountries.The investigation of collaborative mechanisms for properly sharingpolicy incentives through transfer price and value conductiverelationsinthesustainablesupplychaincontextremainssparsein the literature. This is the signicance of this study.2. Literature reviewAsenvironmentalissuescontinueto emerge,research linkingsupplychainmanagement toenvironment issues has attractedattention worldwide. Seuring and Muller (2008) provided anexcellentliteraturereviewonsustainablesupplychainmanage-ment and identied two distinct strategies for sustainable supplychainmanagement: supplier management for risksandperfor-manceandsupplychainmanagement for sustainableproducts.Kosugi et al. (2009) simulatedtheinternalizationof externalitycosts for major global environmental issues and assessedtheresults withalifecycleimpact assessment model. This modelindicatedthat internalizingtheexternalitycost increasesforestpreservation and reduces fossil-fuel consumption. Mitra andWebster(2008)analyzedtheeffectsofgovernmentsubsidiesoncompetitionintheremanufacturingmarket withatwo-periodmodel. They introduced a subsidy scheme that shares the subsidybetween the manufacturer and remanufacturer and increases totalremanufacturing activities. Ding et al. (2014) conducted a study onthe ways to encouragerms to voluntarily produce EFPs to remedynegativeexternalities byfocusingontherelationshipbetweengovernment policies, the economic performance of EFPs, and theimpact ofpolicyincentivesonthecommercialfeasibilityofEFPsusingalifecycleapproach. HolmgrenandAmiri(2007), Nguyen(2008) and Longoa et al. (2008) also conducted a study oninternalizing environmental externality cost from differentperspectives.Studiesof supplychainpricingstrategiesrelatedtoenviron-mental issueslargelyfocusedonreverseorclosedsupplychainproduct pricing decisions. For the resolution of supply chainenvironmental issues, business partnership requires members'mutual cooperationandcoordinationtodecide the retail andwholesale prices, prot margins, and inventory levels to obtain agreatermarket shareandhencehigherrevenues(ChauhanandProth, 2005). Supply chain coordination can be pursued by adopt-ing a centralized or decentralized decision-making approach; theformer option occurs when there is a unique decision maker in thesupply chain and the latter when several independent actors makedecisions at different supply chain stages (Giannoccaro andPontrandolfo, 2004). Vachon and Klassen (2008) conducted astudy that focused on the impact of environmental collaborationin the supply chain on manufacturing and environmental perfor-mance. Their empirical results demonstrated that upstream prac-ticesweremorecloselylinkedwithprocess-basedperformance,whereas downstream collaboration was associated with product-basedperformance. Hsueh(2014) proposeda revenue sharingcontract by embedding corporate social responsibility (RS-CSR) forcoordinating a two-tier supply chain. A mathematical model wasemployed to determine the optimal CSR investment, the wholesaleprice, andtherevenuesharingrationeededtoachievechannelcoordination. The game theory approach is widely used inresearchonthesupplychainpricingproblem. Cai et al. (2009)examined the inuence of price discount contracts on dual-channel supply chain competition. The result showed that aconsistent pricing scheme could decrease channel conictsthrough bringing more prots to retailers. Using Stackelberg gametheory, Yuet al. (2009) discussedhowamanufacturer anditsretailers interact with each other to optimize their individual netprotsbyadjustingproductmarketing(advertisingandpricing)and inventory policies in an information-asymmetric vendormanagedinventory(VMI)supplychain. Cai etal. (2011)studiedoptimal pricingandorderingwithpartiallostsalesfromatwo-stagegametheoryperspective. NagarajanandSoi(2008)sur-veyedapplications of cooperativegametheorytosupplychainmanagement, especiallysupplychainprotallocationandstabi-lity; thestudyalsoreviewednegotiationmodels, includingtheNashbargainingprobleminatypicalnewsvendorsetting, coop-erative bargaining, and negotiation power.Therearenumerousstudiesfocusingontheoperationsman-agement level of greenor reversesupplychains. Tsoulfas andPappis(2006)emphasizedthatenvironmental factorsshouldbeconsideredwhendesigningsupply chainsanddiscussedproductdesign, packaging, transport, recycling, disposal, andgreennessintermsof environmental management. Itwasconsideredthatenvironmental friendliness could be effectively improved throughenvironmental management during the whole life cycle. Chaabaneet al. (2012) introduced mixed integer linear programming, basedon a framework of sustainable supply chain design that considerslife cycle assessment (LCA) principles, in addition to the traditionalmaterial balanceconstraints at eachnodeinthesupplychain.Their results suggested that current legislation and EmissionTradingSchemes(ETSs) must bestrengthenedandharmonizedatthegloballevel to driveameaningfulenvironmental strategy.Mahapatra et al. (2013) proposed a new mode to realize minimumconsumptionof productmaterialsthroughreproduction. Paulraj(2011)empiricallyevaluatedtheeffectof rm-specicresourcesand/orcapabilitiesonsustainablesupplymanagementandsus-tainability performance.Several studieshavearguedthat rmshavenointentiontointernalizetheseexternalities. Morecommonly, rmsarecom-pelled to reduce their impacts on the natural environment by bothregulatoryandnon-regulatorypressures fromthegovernment,market, or evencommunity (Hall, 2000). VermeulenandKok(2012) showed howcompetition between various sustainablesupply chain governance systems in The Netherlands has resultedin recent market breakthroughs, although initial governmentinterventions were very limited and diverse. Chen and Sheu(2009) proposed that well-designedenvironmental pricingstrat-egy regulations can promote extended product responsibility(EPR) for green supply chain members in a competitive market.Mostrelevantstudiesfocusonspecicoperationalaspectsofthe supply chain; however, little research has extended toH. Ding et al. / Int. J. Production Economics () 2Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iinvestigatingsustainablesupplychains fromtheperspectiveofmotivatingsupply chainmemberstocooperatein pollution pre-ventioninvestment decisions. Unlike the current literature, bytakingaviewof assessingeconomicfeasibilityof jointenviron-mental technology investment bysupply chainmembers, thisstudy explores amechanismandpathway forresolving environ-mental externalities using a quantitative approach, which providesacomprehensivemeasureforanalyzingtheenvironmental sus-tainability of supply chains. Specically, we analyze the impact ofsupply chain members'decision-making drivers and governmentpolicy factors on supply chain collaborative pricing decisions, witha focus onpricing strategies that couldrealize the optimizedperformance of sustainable supply chains and unication of supplychain members' satisfaction under the context of different supplychain modes and prot allotment decisions. The remainder of thestudy is organized as follows: Section 3 constructs decentralized,centralized, and collaborative supply chain pricing decision modesbyintroducingEFPs tointernalizeenvironmental externalities;Section 4 presents a case analysis, including discussions; and thenal section presents the conclusions.3. Supply chain pricing model with internalizingenvironmental externality3.1. Model assumptionsThe process of internalization is depicted in Fig. 1, which showsa closed loop with a dotted line that comprises an environmentallysustainable supply chain while considering government regulationand policy incentives. The government's policies promote thesupply chain members to invest in an EFP that replaces an EUFP,which helps reduce negative externalities incurred by the EUFP.The following assumptions are made:1) The supply chain comprises a single product manufacturer andasinglesupplier. Themanufacturer'sproductionof oneunitproduct needs munits of intermediate product fromthesupplier. As a core member of the supply chain, the manufac-turersellsproductsdirectlytoconsumers. Themanufacturerproducesproductsbasedonmake-to-order, andthenishedproducts equal sales. Thesupplier is awareof thedemandinformationof themanufacturer. Themanufacturer andthesupplieradopt avendormanagedinventory(VMI) model tomanagetheirinventory. Thesuppliermanagestheinventorylevel without stock out and bears the inventory cost. Inventoryis replenished in batches on a just-in-time basis, an average of ktimes per year.2) Consideranimperfectcompetitivemarket environment asthe EFP is not the same as the EUFP. Consumer preference likelyhasan impact on pricedifferentials anddemand(salesquan-tity) may increase as the price falls. Consider a pricing strategyfor promotingEFPsales: the product price starts highanddecreaseswithincreasing ofsalesduringitsmarketdiffusionprocess,andthen kept steady after alimitedtimeperiod.Forsimplicity, weassumealineardemandfunctionof price(orviceversa) for theEFPandthat themanufacturer'sproductannual sales quantity Qandprice P are inthe relationofQabP(a40, b40) duringlimitedtimeperiods; that is,whentheEFPreacheseconomicbreakeven, itspriceremainsteady and does not decrease.3) TheproductsproducedaretheEUFPbeforethesupplychainintroduces the EFP, and the production runs at capacity inaccordancewithmarketshare. AftertheEFPisintroduced, theEUFP is gradually replaced within a limited time period and EFPsalesincreasewithgrowthrategthroughitsmarketdiffusion.Supply chain members also seek to maximize their own interestswhen theymaketheir additionalinvestments inenvironmentaltechnologies for pollution reduction and prevention.4) Government provides regulations and incentive-based policiesthat imposepenalties onEUFPs whilesubsidizingEFPs. Weassume that the government grants subsidies directly (or,alternatively, throughconsumers)to themanufacturer, whichmotivates the supplier to collaboratively invest in the EFP. Themanufacturer shares thesubsidywiththesupplier throughtransfer price negotiation.To facilitate our analysis, the following notations are used:Pem; Peim; Pecm Sales price of EFP inthe case of decentralized,centralized, and collaborative decision-making modes, respec-tively (decision variables);Pes; Peis ; PecsSupply chain transfer price of intermediate productfor EFP in the case of decentralized, centralized, and collaborativedecision-makingmodes, respectively;thesearefunctionsof thesupplier's marginal prot (decision variables);Qem; Qeim; Qecm Annual sales quantity of EFPs in the case ofdecentralized, centralized, and collaborative decision-makingmodes, respectively (dependent variables);Bm Policy incentive per unit of output of EFP in proportion toaverageincremental environmental protectioncost (dependentvariable);Qum Average annual demand (market share) of EUFP thatmatches with manufacturer's production capacity;Pum Sales price of EUFP;Pus Supply chain transfer price of intermediate productfor EUFP;Coefcient of policy incentive proportionally related toaverageincremental cost of environmental protectionper unitoutput of manufacturer (0oo1);Policyincentivetransmittedtothesupplier basedonitsaverageincremental cost of environmental protectionper unitoutput;Coefcient of policyincentive transmittedto thesupplierthrough the manufacturer;CAverage incremental environmental protection cost per unitof EFP;Im; Is Project initial investment onpollutionreductionandprevention by manufacturer and supplier, respectively;Cem; Cum Manufacturer's variable operating cost of EFP andEUFP per unit, respectively (excluding procurement cost);Ces; Cus Supplier'svariableoperating cost ofintermediatepro-duct for EFP and EUFP per unit, respectively;Fum; Fus Penalty (non-compliance) cost per unit of outputimposed on EUFP and intermediate product for EUFP, respectively;Ceo; Cuo Supplier'sdeliverycostperbatchquantityofEFPandEUFP, respectively, with VMI mode; Fig. 1. Internalization process of environmental sustainable supply chain.H. Ding et al. / Int. J. Production Economics () 3Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iCeh; Cuh Supplier's inventory holding cost of EFP and EUFP perunit, respectively, with VMI mode;mNumber of units fromsupplier that manufacturer con-sumes to produce onenal product;kAnnual number of delivery times for supplying intermediateproducts to manufacturer, with each batch quantity of mQm/k;gAverage annual growth rate of EFP;n1Time point at which supply chain reaches economic break-even with the EFP;r Risk-adjusted discount rate during initial development per-iod of EFP project;TCSupplier's incremental inventory cost of EFP withVMI mode;NPVm; NPVs; NPVsc Netpresentvalueof EFPprojectformanufacturer, supplier, andsupply chainsystem, respectively, indecentralized decision-making mode;NPVim; NPVis; NPVisc Netpresentvalueof EFPprojectformanufacturer, supplier, andsupply chainsystem, respectively, incentralized decision-making mode;NPVcm; NPVcs; NPVcsc Netpresentvalueof EFPprojectformanufacturer, supplier, andsupply chainsystem, respectively, incollaborative decision-making mode.As mentioned above, as introducing an EFP requires additionalinvestment, supply chain members bear an increased product cost,which leads to a cost disadvantage for the EFP in the competitivemarket; therefore, supply chain members lack motivation to makean environmental protection investment. For the EFP to beeconomicallyfeasible, thegovernmentimposesapenaltyontheEUFP to let it bear the costs of pollution, and offers an incentive-basedpolicytoreducetheproduct cost sothat theEFP'spricecanbecompetitive. Actually, thereasonforthegovernment tohaveanincentivepolicyisthatit hasamajorresponsibilitytoencourage anddrive all business activities inthe directionofenvironmentalprotection ineverypossible way whilecreatingaharmonious relationship between business development andenvironmentalsustainability. Inouranalysis, wetreatproducingandmarketingtheEFPasaproductinvestmentprojectandtheNPV model (a common approach) is used to evaluate the invest-ment project. Government policy incentives aimat offeringmodest compensation for the incremental cost ofthe EFP,whichhelpsthesupplychainsystemachievebreakevenforalimitedtime period at tn1, namely, NPV0. After this point, thegovernment will ceasesubsidies. Accordingtotheprincipleofonly compensating for the investment cost of pollution reductionand prevention, logically the government subsidy should be basedon average incremental pollution prevention cost. As assumed, theamount of policy incentive, denoted by Bm, is offered to themanufacturer to compensate the incremental cost of the EFP. LetCdenotethemanufacturer'saverageincremental pollutionpre-vention unit cost and Bm be measured per unit of the EFP and inproportiontoC. ReferringtoDinget al. (2014), wedenethefollowing:C

n1t 1Cemt CumtmPest PustQemt Im_ _=

n1t 1Qemt1:1Bm C; 0or1 1:2where Qemtis annual sales quantity of the EFP at time period t; Ccomprises additional (initial)investmentcost amortized per unitof the EFP during n1 periods, incremental variable cost per unit ofthe EFP and incremental purchasing cost per unit (transfer price)of the EFP; is the coefcient of policy incentives per unit of theEFP and is negotiable between the government and the manufac-turer. ItsdeterminationdependsonmarketdemandfortheEFP,the investment intensity of innovative environmentaltechnologies, theeffectivenessandefciencyof their usageforenergy saving and pollutant reduction, and business risk; tn1 isthe time point at which the supply chain systemachieveseconomicbreakevenwithNPV0; Pestis thetransfer priceofintermediateproductfortheEFP. Itisassumedthatthemanu-facturer takes an initiative to motivate a sustainable supply chain.As the supplier intends to increase the transfer price to compen-sateforitsincreasedcostduetoproducingtheEFP, themanu-facturerwouldsharethegovernmentsubsidywiththesupplierthrough the transfer price negotiation, i.e., the government policyincentive is transmitted to the supplier through the transfer price.We also assume that the portion of the government subsidyshared by the supplier is proportional to its average annualincremental cost during n1 periods, meaning that the incrementalcost is partially compensated. Rationally, the transfer price of theEFP should comprise the EUFP's transfer price plus the supplier'sincremental cost, wherethelatterincludesincremental variablecost, incrementalinventory cost, andinitialinvestmentcost. Forsimplifying the calculation,we takeCem,Cum,Pes,andPusas annualmean values of Cemt, Cumt, Pest, and Pust, respectively, and rewrite Eq.(1.1) as follows:C Cem Cum mPes Pus Im=

n1t 1Qemt2Ourstudywillfocusonhow apolicyincentivecanaffectthepricingstrategiesof themanufacturerandthesupplier, anditsimpact on the supply chain's value conduction relationship, whichhas a signicance that differs from previous studies.3.2. Environmentally sustainable supply chain model analysisThis section discusses environmentally sustainable supplychainpricingstrategiesof themanufacturerandthesupplierinthe cases of decentralized, centralized, and collaborative decision-making modes while introducing the EFP. With the considerationthat economic feasibility of investment in environmental technol-ogy is mainly reected during early periods of the EFP project, wewill only focus our analysis on the initial development stage (up toachieving breakeven) of the EFP.3.2.1. Supply chain pricing strategy analysis of decentralizeddecision-making modeInthecaseof thedecentralizeddecision-makingmode, themanufacturer andthesupplier eachdevelop theirpricingstrate-gies based on maximizing their own prots. As we treat introdu-cing the EFP as a product investment project involving by both themanufacturerandthesupplier, weevaluatetheirnetprotsbyemployingacashowapproachusingnetpresentvalue(NPV).NPV measures the incremental cash owof the EFP project(comparedwiththeEUFP); here, thenet protof thereplacedEUFP is the opportunity cost of the EFP. The net cashows of themanufacturerof theEFPandtheEUFP, denotedbyRemandRum,respectively, can be expressed as follows:Rem

n1t 1Pemt mPest Cemt Bmt_ _Qemt

n1t 1Pumt mPust_Cemt Fmt_Qum Qemt Im3:1Rum

n1t 1Pumt mPust Cemt Fmt_ _Qum3:2where, as shown in Eq. (3.1), we assume the government subsidyonlygoestothemanufacturer. Noticealsothat weassumetheEUFP is gradually replaced by the same amount as the EFP duringthe initial development stage for QemtrQum. To capture the impactof opportunity cost while replacing the EUFP, by taking differenceH. Ding et al. / Int. J. Production Economics () 4Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iRemt Rumt, the NPV of the manufacturer is expressed as follows:NPVm

n1t 1Pemt Pumt_ _mPest Pust_ _Cemt Cumt_ _Bmt Fumt_ QemtertIm3:3where the manufacturer's product price and the transfer price aredecision variables. In a similar way, we can also derive the NPV ofthe supplier, including the opportunity cost of replacing the EUFP,as follows:NPVs

n1t 1mPest Pust Cest CustFustQemt TCt_ _ertIs3:4where we assume the supplier manages the inventory by means ofVMI (meaning that it bears the inventory cost), TCtis thesupplier's incremental inventory cost associatedwiththe EFP.WithVMImode, themanufacturer andthesupplierjointly setatarget of a minimum inventory cost based on market demand andthemanufacturer'sinventoryismanagedbythesupplierinanagreed commonframework. Consideringthatinventory orderingand holding costs remain unchanged, the total incremental inven-tory cost of the EFP is expressed as follows:TCt kCeo mQemtCeh=2kkCuo mQemtCuh=2k_ _kCeo Cuo_ _mQemtCeh Cuh_ _=2k 4where k denotes the annual number of delivery times for supply-ing intermediate products, kCeo Cuo_ _ is the incremental orderingcost, andmQemtCeh Cuh_ _=2kistheincremental holdingcost. Asassumed, theEFPgraduallyreplacestheEUFPwithanaverageannual growthrategthroughamarket diffusionprocess; it isfurther assumed that the sales quantities for two consecutive timeperiods follow the relation Qmt Qmt 11g. The sales quantityrelations during n1 periods areQmt Qm11gt 1

nt 1Qmt Qm11gn11=g___5Inthe decentralizedcase, the transfer price is determinedwithout cooperationof themanufacturerandthesupplier, andtheprocess maybedividedintotwophases. As product pricedependson market demand, we considerthat depending onthemarket demand the manufacturer determines the EFP's (nalproduct) salespriceor sales quantityfor agiven(preliminary)transfer price in Phase 1. This is followed by the supplierdeterminingitsoptimaltransferprice(only optimalforitself)inPhase 2 based on available information of the EFP's sales price andquantity. WeimplicitlyassumethatthemanufacturertakestheinitiativetomotivatetheEFPthroughthesupply chain. Inthesetwo phases, the backward induction method could be used. In thefollowing part, the optimal pricing strategies of the manufacturerandthesupplierunderthecontext ofthedecentralizedcaseareanalyzed by the backward induction method.(1)Intherstphase, assumingagivenvalueof thetransferprice Pes, the manufacturer determines the optimal price of the EFPwith this given supplier's pricing strategy. On the substitution ofEqs. (1.2) and (2) into Eq. (3.3), we obtain the NPV of themanufacturer as follows:NPVm

n1t 1Pemt Pum m 1 Pes Pus_ _1 Cem Cum_ _ _Im=

n1t 1 Qemt Fum_QemtertIm6On the substitution of the product sales quantity and sales pricerelationQabPandEq. (5) intoEq. (6) andmaximizingthemanufacturer'sNPVby having NPVm=Pemt 0and, for easeofnotation, let M a=bPum m1Pus 1 Cem Cum_ _Fum, wecan obtain the EFP's preliminary price as follows:Pemt a=bm1Pes=2 1g t 17Bysubstituting Eq. (7) backintoQabP, we canfurtherobtain the EFP's preliminary annual sales quantity as follows:Qem1 bm1Pes=2 8(2) Inthesecondphase, themanufacturer shares theEFP'sdemandinformationwiththesupplierandthesupplieraccord-inglydeterminesthetransferpricebymaximizingitsNPV. Sub-stitutingEq.(8) into Eq.(3.4)we canhave thesupplier'sNPV asfollows:Takingthesecond-orderderivativeofEq. (9)withrespecttothe transfer price, we obtain 2NPVs=2Pes bm21o0,whichmeansNPVsisconcaveandhasitsmaximumvalue. Byhaving NPVs=Pest 0 we then obtain the optimal transfer pricein the decentralized case as follows:PensM2m1Pus Ces Cus Fus2Ceh Cuh4k10Noticingthat for maximizingNPVs, thevalueof k(annualinventoryreplenishmenttimes)isgivenandhasnotbeenjusti-ed: this issuewill bedealt withlater inthis section. Onthesubstitution of Pensinto Eqs. (7) and (8), we can obtain the optimalsales price and annual sales quantity of the EFP based on optimalpricing strategy of the supplier as follows:Penmt abM4 m1Pus Ces Cus Fus 4m1 Ceh Cuh8k_ _ 1g t 111Qenm1 bM4bm1Pus Ces Cus Fus4Ceh Cuh_ _8k_ _12When the prices and quantities for both the manufacturer andthesupplierareknown, wecanthendeterminetheamountofgovernment subsidyneededtoreachbreakevenof thesupplychainsystemaccording to Eqs. (1.2) and (2). To simplify thecalculation, we take the mean values of the cost and priceparameters. On the substitution of Eqs. (10)(13) into Eqs.(2) and(1.2), andEqs. (3.3)(3.4), respectively, we obtaintheoptimalamount ofthegovernment subsidyforthe EFP per unit,the maximized NPVs for the manufacturer and the supplier in thedecentralizedcasearedenotedbyNPVnmandNPVns, respec-tively, as follows:Bnm Cem Cum mPensPus Im=Qenm11gn11=g_ 13NPVnm

n1t 1Penmt Pumt_ _mPensPus_ _Cem Cum_ _Bnm_FumtQenm11gt 1ertIm14:1NPVs

n2t 1mb2 Pest Pust Cest CustFustm1Pest 1g t 1kCeo Cuo_ _mbm1Pest Ceh Cuh 4k1g t 1______ertIs9H. Ding et al. / Int. J. Production Economics () 5Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iNPVns

n1t 1mPensPus Ces Cus Fust Ceh Cuh=2kQenm11_gt 1kCeo Cuo_ertIs14:23.2.2. Supply chain pricing strategy analysis of centralized decision-making modeInthecentralizedcase, overallNPVofthesustainablesupplychain system, including the opportunity cost of replacing the EUFP,can be expressed as follows:NPVisc

n1t 1Pemt Pumt_ _mCest Cust_ _Cemt Cumt_ _ _Bmt Fumt mFustQemtert

n1t 1kCeo Cuo_ _mQemtCeh Cuh_ _=2k_ ertIs Im15It is notedthat, different fromthetraditional supplychainmodel found in the literature, when considering the policyincentive, the transfer price exists as the decision variable in thecentralizedmode as above, whichexplicitly presents that thepolicy incentive shared by the supplier for promoting a sustainablesupplychainistransmittedviathetransferprice. Onequestionthat has to be addressed is howto properly distribute thegovernment subsidy between the manufacturer and supplierthrough the transfer price. On the substitution of Eq. (1.2)(representing the government subsidy) and (2), Eq.(4) (represent-ingtheincremental inventorycost) andQabPintoEq. (15),then having NPVisc=Pemt 0, we can obtain the optimal price oftheEFPthatmaximizestheNPVof thesupplychainsystemasfollows:Peinmt ab12abPum mCes Cus_ _1 Cem Cum_ _mPes Pus_ __Fum mFus mCeh Cuh_ _2k_1gt 116WithQabP, wethenobtaintheoptimal salesquantityofthe EFP as follows:Qeinm1 b2abPum mCes Cus_ _1 Cem Cum_ _mPes Pus_ _Fum_mFus mCeh Cuh_ _2k_17SubstitutingEqs. (16)and(17)intoEqs. (3.3)(3.4)and(15),respectively, weobtaintheNPVsforthemanufacturer, supplier,and supply chain system in the centralized case as follows:NPVinm

n1t 1Peinmt Pumt_ _mPes Pus_ _Cem Cum_ __Binm Fumt_Qeinm11gt 1ertIm18:1NPVins

n1t 1mPes Pus Ces Cus Fust_Ceh Cuh=2kQeinm11gt 1kCeo Cuo_ertIs18:2NPVinsc

n1t 1Peinmt Pumt_ _mCest Cust_ _Cemt Cumt_ __ _Binmt Fumt mFust mCeh Cuh_ _=2k_Qeinm11gt 1kCeo Cuo_ __ertIs Im18:3where Binm is determined by Eqs. (1.2) and (2) and based on PeinmtandQeinm1. We see from Eqs. (16) and (17) and Eqs. (18.1)(18.3) that Peinmt,Qeinm1, the NPVs of the manufacturer, supplier, and supply chain, alldepend on the transfer price Pesthat incorporates the governmentpolicyincentive; therefore, weneedtodeterminetheoptimalvalues of sales price and quantity of the EFP through justifying thetransfer price. One might consider working out an optimal transferpricethat maximizes NPVinsc; however, bytakingthesecond-order derivative of Eq. (18.3) with respect to the transfer price wehave2NPVinsc=2Pes bm2=240 18:4The above means that NPVinscis a convexfunctionof thetransfer price and it cannot be maximized through itsrst-orderderivative condition NPVinsc=Pes 0; thus, we then need toemploy an alternative approach tond the optimal transfer pricethat ensures a higher NPV of the EFP project for the supply chainthan the decentralized case and simultaneously satises both themanufacturer and the supplier. As assumed, the governmentsubsidy is only provided to the manufacturer and there is a needtoproperly allotthegovernmentsubsidybetweenthemembersthroughcollaborativenegotiation, whichisalsoconnectedwiththe transfer price determination. This means that the central issueis the transfer price negotiation, including allotment of thegovernment subsidy between the manufacturer and the supplier,on which the determination of the optimal price and salesquantityof the EFPis based. Inthe followingsections, as analternative approach, Rubinstein game theory (Rubinstein, 1982) isadopted for obtaining the optimal transfer price in our model; thatis, we conduct a collaborative decision process of the transfer pricethrough negotiation between the members.3.2.3. Supply chain pricing strategy analysis of the collaborativedecision-making modeLogically, optimal sales quantity Qeinmtin the centralized case isnormally larger than Qenmtin the decentralized case, which meansthat there exists room for more sales, more potential prots, andpossibly resulting in a higher NPVinsc; however, the way of sharingincreasednet prots tosatisfyboththemanufacturer andthesupplierusuallyoccursthroughnegotiation. Thatis, sharingtheincreased net prots gained from the centralized supply chain in acollaborativemannerdependsonthetransferpricenegotiationbetweenthesemembers. Toclarifytheinterconnectionbetweenthe transfer price and government policy incentive, we dene thetransfer price for the EFP as follows:Pes Pus

n1t 1Cest Cust mQemt TCt Is_ _=m

n1t 1 QemtZ019:1where is the policy incentive transmitted to the supplier basedonitsaverageincremental costofenvironmental protectionperunit output. As seen in Eq. (19.1), the transfer price Pesdepends ontheaverageannual incremental cost. Government policyincen-tives change supply chain's value transition and affect the supplychain's pricing strategy, the impact of which is mainly reected intwoaspects: rst, compensationforpart oftheincrementalcostwould have the supplier not price the intermediate product for theEFPtoohigh, whichweakensthecostdisadvantageof theEFP;second, the manufacturer and the supplier negotiate the allotmentof thegovernment subsidythroughadjustment of thetransferprice. Onthesubstitutionof Eq. (19.1) intoEq. (2), werewriteEq. (1.2) as follows:Bm _Cem Cum Im=

n1t 1 Qemt_m_Ces Cus

n1t 1TCt_ _Is_=m

n1t 1 Qemt_; 0oo1 19:2wherethecost parameterstaketheirannual meanvalues. Theabove expression reects the government subsidy offered tocompensate the supply chain's incremental costs for the EFP: therst term on the right-hand side is to the manufacturer, the secondtermistothesupplier. Sincethemanufacturermaynotobtainenough subsidy from the government that could fully compensateits incremental cost (o1), o1 means the manufacturer wouldH. Ding et al. / Int. J. Production Economics () 6Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iunlikelycompensatethesupplier'sincremental costfully. Inthecollaborative case, based on its incremental cost, the manufacturernegotiates the transfer price with the supplier through adjustmentof the proportion of subsidy shared with the latter and the nego-tiation result should satisfy the following conditions:NPVcmZNPVnm; NPVcsZNPVns20:1where NPVcmand NPVcsdenote feasible NPVs of the EFP projectfor the manufacturer and the supplier in the collaborative case, andNPVnm and NPVnsare the decentralized optimal NPVs, respectively.The above inequalities simply mean that in the collaborative case, netprots shared by each of individual supply chain member should notbelessthanthatearnedinthedecentralizedcase. TheconditionsNPVcsZNPVnsand NPVcmZNPVnm, respectively, determine thelower and upper bounds of the transfer price that are feasible for acollaboration. Since solving the above two inequalities for the lowerand upper bounds of the transfer price is very complex by nature (i.e.,lessmeaningful inapractical sense), weturntoamorefeasiblealternative method to determine them. Noticing from Eq. (19.1) thatthe transfer price depends on the coefcient and the supplier's inc-remental cost, the lower bound of the transfer price can be expressedas follows:Pes min Pusand Qes 0 for 0 20:2Compared with the decentralized case, logically the centralizedoptimal salesquantityQeinmt4Qenmtisnormallylargerandcorre-sponds to a lower transfer price Peins oPens ; thus, the upper boundof the transfer price is as follows:Pes max Pens20:3Eqs. (20.2) and (20.3) mean that in the collaborative case, thetransferpricetakesitsminimum(lowerbound)whennoEFPisproducedwithout agovernment policyincentive, andtakesitsmaximum(upper bound), whichcorresponds tothesupplier'smaximized interests, in the decentralized case. We can then inferthatPes minoPesoPes max. However, theselowerandupperboundsmay not satisfy Eq. (20.1) and, therefore, might not be valid.NoticingfromEq. (18.4) that theNPVof theintegratedsupplychain NPViscis convex, we can show that the transfer price thatcorresponds to its minimum value, denoted by Pesl, is less than Pus(see Appendix), i.e., PesloPus Pes minoPesoPes max, which means thevalid lower and upper bounds of the transfer price must be alongthe right part of the NPVisc curve that moves up. Noticing that forother given parameters Eqs. (16) and (17) and Eqs. (18.1)(18.3) arethe general functions of the transfer price for an integrated supplychain, for the collaborative case, as the transfer price PecsincreasesNPVcscwillalsoincrease;correspondingly, NPVcsincreases andNPVcmdecreases, respectively, as canbe shownfor NPVcs=Pes40andNPVcm=Peso0forPes4Pus. Itfollowsthattherearevalid lower and upper bounds, denoted by Pecs min and Pecs max, whichcorrespond to NPVcsZNPVns(for PusoPecs minoPecs) andNPVcmZNPVnm(forPecsoPecs maxoPes max), respectively. Wethenhave:Pecs max4Pecns4Pecs min20:4where Pecnsis the collaborative optimal transfer price. Heuristically,we can obtainPecs minandPecs maxby examining theincrease in thetransferpricefollowedbyndingtheoptimaltransferpricePecnsvia collaborative negotiation through the following logic:1) Noticing that both Pecmt and Qecm1 will change as the transfer pricechanges, inturn, NPVcsc; NPVcs and NPVcmwill changeaswell. Forthecollaborativecase, NPVcsincreasesandNPVcmdecreases, respectively, asthetransferpriceincreases(ascanbe shown for NPVcs=Pes40 and NPVcm=Peso0 for Pes4Pus).2) Assumethetransferpriceincreases(startingfromPes min) ordecreases (starting from Pes max). In the former case, NPVcswillincreaseuptothelevel NPVcs NPVnsthatcorrespondstothe lower bound Pecs min; in the latter case, NPVcm will increaseup to the level NPVcm NPVnm that corresponds to the upperbound Pecs max. The processes above will be explained further inFigs. 2 and 3 in Section 4.3) [Pecs min, Pecs max] is thetransfer priceinterval that satises Eq.(20.1) and is feasible for the collaborative case.4) Within this feasible interval, when the transfer price increasesup to PecsPecs max, the value of NPVisc(or NPVcsc) reaches itsmaximum; however, thisisnotoptimal since, insuchasit-uation, themost increasednet protsgainedfromtheinte-grated supply chainaretakenby thesupplierandthemanu-facturer would not be happy with just the same as thedecentralized case (no gain from collaboration).5) Therefore, for a collaboration to be successful, the manufacturerandsuppliermust eachbesatisedwiththeirprotssharingviathetransferpricenegotiations. Inourmodel, theoptimaltransfer pricePecnsis obtainedthroughtheRubinsteingameapproach and, in turn, the optimal price Pecnmtand sales quantityQecnm1are determined based on Pecns.The collaborative process is a bargaining process to allocate thesupply chain benet between the members in a proper manner. Byadopting the Rubinstein game approach to determine negotiatedtransfer price in an indenite bargain game with two participants,the unique sub-game perfect Nash equilibrium, namely, the opti-mum of the transfer price can be achieved. The ratio of the supplychainbenet allotment betweenthemembers intermsof thetransfer price can then be expressed as follows:Pecs max PecnsPecnsPecs min1221121:1Rewriting the above equality yields:Pecns211Pecs max 12 Pecs min11221:2where 1and 2(0rr1) represent the discounting factors(degree of patience) of the manufacturer and the supplier, respec-tively. Assume the manufacturer and the supplier bear axed costfor each bargaining period and each of them has axed discount-ing factor that can be considered as the cost of bargain. The majorinuential factors for the degree of patience include marketposition, core competence, utility function, negotiation cost,degree of risk aversion, etc. Inthe collaborative case, onthesubstitutionofPecnsinto Eqs. (16)and(17)andEqs. (18.1)(18.3),we obtain the EFP's optimal sales price Pecnmand sales quantity Qecnm,and the NPVs of the EFP project for the manufacturer, supplier, andsupply chain system.3.2.4. Justication of VMI policyIntheaboveanalysisofoptimizingtheEFPprojecttorealizebreakeven, the number of supplier delivery times k is given, whichmay not meantheinventorymanagement policyis optimal.Theinventory management policy can be optimized through adjustingk and by doing so we can also examine whether inventorymanagement policy is sensitive to the EFP sales quantity requiredforbreakeven. Themethodadoptedhereistoset avalueof kinitially, then solve for an optimal sales quantity Qeinmtthat satisesboth the collective and individual interests according to Eqs. (16)and (17). Next, we use an economic ordering quantity (EOQ) modeltodeterminethevalueof k, denotedbykn, thatcorrespondstoordering cost equal to holding cost (recognizing that for the EOQH. Ding et al. / Int. J. Production Economics () 7Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016imodel, the total inventory cost is at the minimum when orderingcost equals holding cost):kCeo Cuo mQeinmtCeh Cuh=2k; knmQeinmtCeh Cuh=2Ceo Cuo 22Alternatively, we may regard k as a variable and letNPVisc=k 0; thus, kncan be derived as above. On the substitu-tion of kninto Eqs. (16), (17) and (21), we redo the optimization ofthetransferprice, salesprice, andsalesquantityoftheEFP. WeFig. 2. Lower bound of the transfer price.Fig. 3. Upper bound of the transfer price.H. Ding et al. / Int. J. Production Economics () 8Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016ithenfurtherrepeattheoptimizationprocesstoobtainthenewvalue of knuntil the difference between two adjacent values of knis less than a certain given error. Up to this point, knis the optimalinventory management policy and the corresponding values of thetransfer price, sales price, and sales quantity of the EFP are optimalin the centralized model.3.2.5. Impact of the government policy incentive on pricing strategiesAs described above, government policy incentives compensatefor the incremental cost associated with producing the EFP; thus,it has an impact on the supply chain members' pricing strategies.To further analyze the relations of coefcients of policy incentive with the transfer price, sales price, and sales quantity of the EFP,forEqs. (10)(12), (16)and(17), wetaketheir rst-orderpartialderivatives with respect to , respectively, as follows:Penst = a=bPum Fum2m1240Penmt= _Cem CummCes Cus Fus 4mCeh Cuh8k_ 1g t 1o0Qenm1= bCem Cum4bm_Ces Cus Fus4Ceh Cuh8k_40Peinmt= _Cem CummPes Pus _1gt 1=2o0Qeinm1= b_Cem Cum mPes Pus _=240Asseenintheabove, inbothdecentralizedandcentralizedcases, byincreasingcoefcient , thesalespriceof theEFPwilldecrease and thus its sales quantity will increase. This is consistentwith the government's intention of providing a subsidy to enlargethe EFP's market proportion by reducing its cost. The transfer priceincreases with increasing, which implies that the manufacturershares the government subsidy with the supplier. In analyzing Eqs.(14.1) and (14.2), we can also see that with increasing , the NPVsof both the manufacturer and the supplier increase.In a similar way, we can further analyze the impact of penaltycost on the pricing strategies and performance of investing in theEFP. Bytakingthe rst-orderpartial derivativesof thetransferprice, salesprice, andsalesquantitywithrespecttoFumandFus,respectively, we obtain:Pens=Fus 1=2o0Penmt=Fum 1g t 1=4o0Qenm1=Fum b=440Peinmt=Fum 1gt 1=2o0Qeinm1=Fum b=240Fromthe above we can see that the EFP's sales quantityincreases withpenaltycost, whichimplies that morerigorouslegislationdrivesthetendencytoproducemoreEFPs, whichinturn lowers both the sales price and transfer price.4. Case analysis and discussionsInthissection, we presentacaseanalysisofahybridvehicleprojectfromanautocompanytoquantifytheimpactof policyincentives onthe environmental performance of a sustainablesupply chain.4.1. Data collectionTherelateddataareestimated basedonthemarket situationshown in Table 1.4.2. Numerical analysisThe analytical results for the data presented above are showninTable2. Theprocess of searchingfor thelower andupperbounds of the transfer price are shown in Figs. 2 and 3.In the centralized case, NPVcs(or NPVis) is convex andNPVcm(or NPVim) is concave with respect to the transferpricefor Pes4Pus(2NPVcs=Pes2bm2and2NPVcm=Pes2bm21=2); thus, for a valid transfer price, NPVcm decreasesand NPVcsincreases withPes.Figs.2and 3,respectively,presentthe determination of the lower and upper bounds of the transferprice, where the dotted curves, respectively, represent NPVs andNPVm in the decentralized case and the solid curves, respectively,represent NPVcs (or NPVis) and NPVcm (or NPVim), respectively,in the centralized case. As seen in Fig. 2, the NPVs (decentralized)takes its maximumat point A1(Pens144.76) with NPVns 1062.47. We can ndpoint B1at whichNPVcs NPVnswithPes 102.09. When Pes4102.09, NPVcs4NPVns, i.e., Pes min 102.09is the lower bound of the transfer price. As shown in Fig. 3, at pointA2 we have Pens144.76 and NPVnm 712.44; correspondingly, wecan ndpoint B2at whichNPVcm NPVnm, withPes 113.94.When Peso113.94, NPVcm4NPVnm, i.e., Pes max 113.94 is theupper boundof thetransfer price. Fig. 4showstheprocessofdetermining the lower and upper bounds of the collaborative tra-nsfer price, where the shadow part on the horizontal axis showsthe feasible collaborative transfer price interval [Pecs min, Pecs max].Table 1Initial investment cost and operating data.Pollution prevention initial investment costitemsMillionInvestment of manufacturer (Im) 4000Investment of supplier (Is) 3000Operating data itemsItemThousand/UnitItemThousand/UnitItem Item timesPum120 Cem50 0.15 k (initialvalue)48Pus60 Cum40 10.1 m 1Fum5 Ces30 20.3 Item QuantityFus4 Cus20 r 10% Qem210; 0000:72PemCeh0.15 Cuh0.1 g 15% Qum100,000Ceo0.65 Cuo0.6Table 2Optimal valuesof Qem; Pem; PesandNPV: comparisonof decentralizedandcolla-borative pricing strategies ( n13 years).Item (thousand/unit)Decentralized Collaborative Justication of VMI policy(kn192 times)C 65.41 43.00 42.99Bm 9.81 6.46 6.45Pes144.76 105.39 105.39Pemn1218.23 170.00 170.00Qem1(thousandvehicles)26.78 55.92 55.92Qemn1(thousandvehicles)35.42 73.96 73.96 0.51 0.27 0.27NPVm(million) 712.44 1546.92 1546.92NPVs (million) 1062.47 1417.55 1417.59NPVsc(million) 1774.91 2964.47 2964.52H. Ding et al. / Int. J. Production Economics () 9Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iNoticethattherearethreeNPVbreakevenscorrespondingtothe manufacturer, the supplier, and the supply chain system,respectively; however, for a given value of the transfer price theymay not happen at the same time. We dene n1 as the time pointat which the latest breakeven is realized, which is considered thesupply chainsystembreakeveninthenumericalanalysis. Attheend of the initial development stage (tn1), the manufacturer, thesupplier, and the supply chain with the hybrid vehicle project willall makebreakeven, i.e., theirNPVsbecomenonnegative. Theseresults are shown in Table 2, with n13 years for comparison.As shown in Table 2, compared with the decentralized case, thecollaborativecaseisbetteroverall. Inthecollaborativecase, theaverageincremental cost andgovernment subsidyareless, thesales quantityof theEFPis larger, andthesales priceis less;readjusting the transfer price creates a higher overall NPV of theEFPprojectforthesupplychainandalsoimprovesthebenetsgained by each of its members. After breakeven, the transfer pricebecomes Pes Pus Ces Cus , including only the increment variablecost per unit of output.From Eq. (22), we obtain kn192 times, meaning that inventoryshouldbedeliveredaroundeverytwodays. Followingupthesubstitutionofkn192into Eqs. (16), (17), and(21.2), we obtainnewvalues(Peinmt 170051.95 RMB/unit, Qeinm155923.20vehicles),whichare veryclose tothe results whenk48times (Peinmt 170052.19 RMB/unit, Qeinm155923.06vehicles), showingthat thepricingstrategyofthesupplychainmaynotbesensitivetotheinventoryreplenishment timescalculatedwiththeVMI model.This may infer that when inventory delivery and holding costs arerelativelysmall, theNPVof thesupplychainislesssensitivetoinventoryreplenishmenttimes. WiththeVMI mode, themanu-facturerandthesuppliercansettletheinventoryreplenishmenttimes and quantities suitable to both sides.AsshowninTable3, ithappensthatunderboththedecen-tralized and collaborative models, n13 years (for all of the threeNPVs to become positive).In reality, n1 may differ with differentpricing strategies and it likely becomes shorter in the collaborativecase, showing that collaboration between members improves thesupplychain'sperformance. Weseethat collaborationnotonlyimproves integrated supply chain performance but also the mem-bers' performances.The discountingfactors are inuencedbythesupplychainmembers' market positions, core capacities, bargaining costs, riskpreferences, etc.; inpractice, thesecanbeestimatedbasedonsupply chainmembers' experiences. Wecanshowtheimpactofchangingbargainingpositiononthetransferprice. Forinstance,when changing the discounting factor of the supplier from 2 0.2to0.3whilekeeping10.1, theoptimal transferpricePecnswillincrease from 104265.60 RMB/unit to 105387.73 RMB/unit; on theotherhand, whenchangingthediscountingfactorofthemanu-facturer from1 0.2 to 0.3 while keeping 2 0.1, Pecnswilldecrease from 103056.19 RMB/unit down to 102943.98 RMB/unit,which implies that with the supplier's bargain position becomingbetterthetransferpriceincreases, whereaswhenthemanufac-turer's bargain position gets better the transfer price decreases.The impacts of the policy incentive and penalty cost on pricingstrategies and supply chain performance in the collaborative modeFig. 4. Process of determining the lower and upper bounds of the collaborative transfer price.Table 3Comparison of NPV for decentralized and collaborative pricing strategies.Decentralized CollaborativeM S SC M S SCt1 2645.94 1091.42 3737.37 1912.92 1006.96 2919.88t2 1665.74 894.57 771.17 595.45 1066.93 471.47t3n1712.44 1062.47 1774.92 1546.93 1417.55 2964.48t4 848.95 1237.19 2086.14 1831.99 1782.40 3614.38t5 991.00 1418.99 2409.99 2128.61 2162.04 4290.65Mmanufacturer, Ssupplier and SCsupply chain.H. Ding et al. / Int. J. Production Economics () 10Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016ican be tested through sensitivity analysis. The results are shown inTables 4 and 5, respectively.4.3. DiscussionFromthe results showninTable 4, we cansee that withincreasing the government subsidy ratio , the EFP's sales quantityincreases as thesales pricedecreases; meanwhile, thetransferprice increases. This implies that the more the incremental cost iscompensatedbytheincreasedgovernmentsubsidy, thelessthecost disadvantage to the manufacturer. The manufacturer then hasmoreroomtolower thesales pricewhileincreasingthesalesquantity, as long as its total prot can be increased. The fact thatthe transfer price increases with meansthat more incrementalcosts of the supplier are compensated by sharing the governmentsubsidy. As shown in Table 5, with increasing the manufacturer'spenaltycost, theEFP'ssalesquantityincreases, salespricedecr-eases, the government subsidy decreases, and, similarly, thetransferpricedecreaseswiththesupplier'spenaltycostaswell.This is intuitively true since the more penalty cost on the EUFP, themore EFPs are preferred to replace the EUFP and consequently lessgovernment subsidy is needed. The NPVs of the members and thesupply chain increase because of increasing EFP sales quantity.As a focused issue connected with the effect of the governmentpolicies on promoting environmentally friendly activities, ourstudyndsthatthepreferentialpolicyincentivepositivelyimp-actsEFPprojectperformance. Theimplicationisthatduetothecrucialroleofsupply chainmembersinreducingenvironmentalexternalities, itisessential forthegovernment, byofferingeffe-ctiveincentives, tohavethesupplychainmembersvoluntarilyaccept their environmental responsibilities. While readjusting thetransfer price to balance the supply chain benet allotmentbetween members, the distribution of the government subsidy islogically interrelated with the transfer price. Our results show thatthe transfer price of the collaborative case is lower than that of thedecentralized case, which means that in the former, it makes senseforthesupplierto share thegovernmentsubsidyto compensatefor its increased cost of preventing environmental pollution whilereducingthetendencytooverchargethepurchasingcostof themanufacturer.In our model, the transfer price presented in the centralized caseinterrelates to the policy incentive and affects the supply chain mem-bers' payoffs. Onemayconsiderthatthetransferpriceshouldberegardedas adecisionvariableandits optimal valueshouldbeobtainedinthecentralizeddecision-makingmodel. However, ourresultscomeoutwiththenontrivial ndingsthattheNPVof theintegrated supply chain is convex with respect to the transfer price.This means the optimal transfer price cannot be worked out througha conventional optimality conditions method; therefore, we have toseek an alternative approach to a collaborative solution. The implica-tion is that within a supply chain context, when introducing a policyincentive in terms of compensating extra cost, the members mightface a complex problem connected with prot allotment, and morelikelythememberswill workout asatisfactorysolutionthroughcollaborative negotiation. In our study, we present a heuristic methodto obtainanegotiatedcollaborativesolutionthatis satisfactoryforthe members through nding out the lower and upper bounds of thefeasibletransfer pricetogether byadoptingtheRubinsteingameapproach.Inthecontext of asupplychain, anoften-mentionedissuerelates to howmuch the government will offer as a policyincentive to compensate supply chain members' incremental costsassociatedwithproducingtheEFP. Inour model, thevalueofcoefcientisassumedas(o1), meaningthatisnegotiatedbetweenthegovernmentagencyandthemanufacturerandtheincremental costof themanufacturerwill notbefullycompen-sated. Theimplicationisthatthegovernmentagencyrealizesamoralhazardeffectthatmightbecreatedif themanufacturerclaims a disproportionate incremental cost; on the other hand, theintention of the government's policy incentive is actually topromote the supply chain members to take the initiative toimprove their environmental performances through technologicalinnovation; consequently, such initiatives enable the supply chaintomovetowardsitseconomicbreakevenbygainingconsumers'acceptance in environmental interests Related to this aspect,a connectedissue arises regarding the factors that shouldbeTable 4Impact of the policy incentive on pricing strategies and NPV in the collaborative mode.Government subsidy rate Collaborative model for n13Bm(thousand) Pes(thousand) Pem(thousand) Qem(unit) NPVm(million) NPVs (million) NPVsc(million)0.10 4.34 103.74 171.99 72405.91 1199.88 1157.96 2357.840.15 6.45 105.39 170.00 73958.24 1546.93 1417.55 2964.480.20 8.51 107.18 167.98 75612.52 1884.66 1705.21 3589.870.25 10.52 109.13 165.77 77382.64 2210.66 2025.96 4236.62Table 5Impact of penalty costs on pricing strategies and NPV in the collaborative mode.Penalty cost (thousand/unit) Collaborative model for n13Bm(thousand) Pem(thousand) Qem(unit) NPVm(million) NPVs (million) NPVsc(million)Fum 5 6.45 170.00 73958.24 1546.93 1417.55 2964.48Fum 6 6.43 169.36 74511.57 1616.50 1482.50 3099.00Fum 7 6.40 168.67 75064.90 1686.60 1547.92 3234.52Fum 8 6.38 167.98 75618.23 1757.22 1613.81 3371.04Pes(thousand) NPVm(million) NPVs (million) NPVsc(million)Fus 4 105.39 1546.93 1417.55 2964.48Fus 5 104.39 1584.5 1499.54 3084.04Fus 6 103.39 1622.19 1582.21 3204.40Fus 7 102.39 1659.98 1665.56 3325.54H. Ding et al. / Int. J. Production Economics () 11Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iconsidered for determining the transfer price through negotiationbetweenthemembers. Fromthemanufacturer's point-of-view,only the supplier's incremental cost can possibly be compensatedpartially(o1), andthecompensatedproportionwill bethefocal issue. However, fromthesupplier side, as shownintheresultsofcaseexampleforthedecentralizedcaseandbasedonmaximizing its own prot, the supplier often intends to charge ahigherpriceby addingmorethanitsincrementalcostontopofthe EUFP's transfer price, which in turn drives the sales price of theEFP higher than expected. The implication is that, as a core issue,the transfer price negotiation should be mediated through an on-goingcollaborationbetweenthememberstoachievesuccessinproducing the EFP.In our model, inventory delivery frequency seems to have littleimpact on operational performance when the VMI model is used.Our numerical results showthat theEFP's sales priceor salesquantitymaynotbesensitivetoinventoryreplenishmenttimeswiththeVMI model. Theimplicationisthatbusinessdecisions,such as product investment projects, are taken at a strategic levelandannual salesquantityandsalespricearedecisionvariablesconcerning economic breakeven. This differsfrom the traditionaloperations management model, wherethesevariables arenor-mally treated as given parameters. For instance, plant-level opera-tional decisionsaremadeundergivenproductsalesordersandprice, and more often concern cost minimization with efciency ofshopoor daily operations, such as inventory control, productionschedule, order quantity, service level, etc.: these are the decisionsmade at the shop oor-level that may have little impact onstrategic level decision making.5. ConclusionsEnvironmental pollution reduction and prevention is the eter-nal topic:theprocessofreducingenvironmental externalitiesismuchmorecomplexinreality. Takingperspectives frombothsupply chain members and the government, our study constructs asustainable supplychainmodel toevaluatethe economic andenvironmental performance of a supply chain. By explicitly intro-ducing regulations and policy incentives in the context of a supplychain, the contribution lies in the investigation of the mechanismfor motivatingrms to collaboratively produce EFPs with environ-mental technology investment, and a quantitative analysis of gov-ernment policies on reducing negative externalities. Based on theprospectofanEFPprojectinvestmentdecision, weestablishanenvironmentallysustainablesupply chain pricingstrategymodelof theEFPsinthecasesofdecentralized, centralized, andcolla-borativedecision-making. Ourstudyndsthat, unlikethetradi-tional supply chain model when considering policy incentives, thetransfer price plays its role in the centralized supply chain mode,whichexplicitlypresents that thegovernment policyincentiveshared between members is transmitted via the transfer price. Ourstudy reaches the following conclusions:1) Withcooperationbetweenmanufacturer andsupplierin pro-moting environmental pollutionpreventioninvestment, theoverall performance of the environmentally sustainable supplychain can be improved and, through further collaborative adj-ustment of the transfer price that incorporates the governmentpolicy incentive, both collective and individual interests can beunied with satisfaction.2) Government incentive policies impact the environmentallysustainable supply chain pricing strategy. A reasonably distrib-uted government subsidy helps the EFP price to adapt tomarket competition with less cost disadvantage, enlarging thesales quantity and increasing not only the overall benet of thesupplychainas awhole but alsoindividual benet of thesupply chain members.3) Duetoincorporatingthegovernmentpolicyincentive, inthecentralizedcasetheNPVofthesupplychainturnsouttobeconvex with respect to the transfer price in our model, so thatwe need an alternative collaborative approach to determine atransfer price that is acceptable for the supply chain members.EmployingtheRubinsteingameapproach, thecollaborationbetween supply chain members can be realized and a satisfac-torytransferpricedeterminedthatwillunifytheinterestsofboth the collective group and individuals.4) Incremental costsof boththemanufacturerandthesupplierassociatedwithproducingtheEFParepartiallycompensatedby the government subsidy through transfer price negotiation,and the motivation mechanism is worked out by realizing theNPV breakeven of the supply chain system. By doing so,investment in the EFP can become economically feasible.There are open issues that remain to be examined. For instance,this study is limited by considering only one manufacturer and onesupplier in a sustainable supply chain, whereas extending this frame-work to a more complex supply chains and other industries wouldimprovegeneralizabilityof theseresultsandexpandtheresearchpotential. Alternatively, the question of whether, without governmentpolicyincentives, supplychainmembers' environmental pollutionprevention investments can jointly make breakeven through innova-tions in the competitive market could be another avenue for futureresearch.AcknowledgmentsTheauthorswouldliketothanktheanonymousrefereesfortheir valuable comments and suggestions to improve the work ofthis paper.This researchis supportedbythe National Natural ScienceFoundation of China with Grant no. 71372012, partially supportedby the Research Fund for the Doctoral Program of High Educationof China with Grant no. 20130009110038, and the Beijing PlanningOfce of Philosophy and Social Science Foundation of China withGrant no. 13JGB021.AppendixFromEq. (15), by having NPVisc=Pes 0, we can obtainPesl Pus A, whereA a=bPum Fum mFus mCes Cus 1 Cem CummCeh Cuh=2k=mSince a=b=Pumis positive and relatively larger, and incremen-tal costs are relatively smaller, we can normally infer A40 so thatPesloPus.ReferencesCai, G.G., Zhang, Z.G., Zhang, M., 2009. Gametheoretical perspectivesondual-channel supply chaincompetition withpricediscountsandpricingschemes.Int. J. Prod. Econ. 117 (1), 8096.Cai, G.G., Chiang, W.C., Chen, X., 2011. Game theoretic pricing and orderingdecisionswith partial lost sales in two-stage supply chains.Int. J. Prod. Econ.130 (2), 175185.Chaabane, A., Ramudhin, A., Paquet, M., 2012. Design of sustainable supply chainsunder the emission trading scheme. Int. J. Prod. Econ. 135 (1), 3749.Chauhan, S.S., Proth, J.M., 2005. Analysis of a supply chain partnership with revenuesharing. Int. J. Prod. Econ. 97 (1), 4451.Chen, Y.J, Sheu, J.B., 2009. Environmental-regulationpricingstrategiesforgreensupply chain management. Transp. Res. E: Logist. Transp. Rev. 45 (5), 667677.Coase, R.H., 1960. The problem of social cost. J. Law Econ. 3, 143.H. Ding et al. / Int. J. Production Economics () 12Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016iDing, H., He, M., Deng, C., 2014. Lifecycleapproachtoassessingenvironmentalfriendly product project with internalizing environmental externality. J. Clean.Prod. 66, 128138.Giannoccaro, I., Pontrandolfo, P., 2004. Supply chaincoordinationby revenuesharing contracts. Int. J. Prod. Econ. 89 (2), 131139.Hall, J., 2000. Environmental supply chain dynamics. J. Clean. Prod. 8 (6), 455471.Holmgren, K., Amiri, S., 2007. Internalisingexternal costsofelectricityandheatproduction in a municipal energy system. Energy Policy 35, 52425253.Hsueh, C.-F., 2014. Improvingcorporatesocial responsibilityinasupplychainthrough a new revenue sharing contract. Int. J. Prod. Econ. 151, 214222.Kovacs, G., 2008. Corporate environmental responsibility inthe supplychain.J. Clean. Prod. 16, 15711578.Kosugi, T., Tokimatsu, K., Kurosawa, A., Itsubo, N., Yagita, H., Sakagami, M., 2009.Internalizationof theexternal costsof global environmental damageinanintegrated assessment model. Energy Policy 37, 26642678.Longoa, A., Markandya, A., Petrucci, M., 2008. The internalization of externalities inthe production of electricity: willingness to pay for the attributes of a policy forrenewable energy. Ecol. Econ. 67, 140152.Mahapatra, R.N., Biswal, B.B., Mohanty, R.P., 2013. A modied reverse supply chainwith remanufacturing for sustainable product cycle. Int. J. Supply Chain Manag.2 (2), 5159.Mitra, S., Webster, S., 2008. Competitioninremanufacturingandtheeffectsofgovernment subsidies. Int. J. Prod. Econ. 111, 287298.Nagarajan, M., Soi, G., 2008. Game-theoretic analysis of cooperationamongsupply chain agents: review and extensions. Eur. J. Oper. Res. 187 (3), 719745.Nguyen, K.Q., 2008. Internalizingexternalitiesintocapacityexpansionplanning:the case of electricity in Vietnam. Energy 33, 740746.Paulraj, A., 2011. Understanding the relationships between internal resources andcapabilities, sustainablesupplymanagementandorganizationsustainability.J. Supply Chain Manag. 47 (1), 1937.Rubinstein, A., 1982. Perfect equilibriumina bargainingmodel. Econometrica50 (1), 97109.Seuring, S., Muller, M., 2008. From a literature review to a conceptual frameworkfor sustainable supply chain management. J. Clean. Prod. 16, 16991710.Tsoulfas, G.T., Pappis, C.P., 2006. Environmental principles applicabletosupplychains design and operation. J. Clean. Prod. 14 (18), 15931602.Vermeulen, W.J.V., Kok, M.T.J., 2012. Government interventions insustainablesupply chain governance: experience in Dutch front-running cases. Ecol. Econ.83, 183196.Vachon, S., Klassen, R.D., 2008. Environmental management andmanufacturingperformance:theroleofcollaborationinthesupply chain. Int. J. Prod. Econ.111, 299315.Yu, Y., Huang, G.Q., Liang, L., 2009. Stackelberg game-theoretic model for optimiz-ingadvertising, pricingandinventorypoliciesinvendormanagedinventory(VMI) production supply chains. Comput. Ind. Eng. 57 (1), 368382.H. Ding et al. / Int. J. Production Economics () 13Please cite this article as: Ding, H., et al., Pricing strategy of environmental sustainable supply chain with internalizing externalities.International Journal of Production Economics (2015), http://dx.doi.org/10.1016/j.ijpe.2015.05.016i