Research Article

Analyzing a socially responsible closed‑loop distribution channel

with recycling facility
Nikunja Mohan Modak1 · Sudipta Sinha2 · Shibaji Panda3 · Nima Kazemi4

© Springer Nature Switzerland AG 2019

Abstract
This paper deals with a closed-loop distribution channel consisting of a socially responsible manufacturer, multiple retail-
ers and a third party collector. In reality, collection of used products (plastic, glass, metal) by a third party collector is more
common than the collection through retailers. This is because retailers generally faces difficulties  such as lack of space
and manpower. Aligned with many closed loop supply chains, this paper assumes that the third party operates the reverse
channel by collecting the used products. The third party collects used products, segregates recyclable items and sends
them to the manufacturer for further use. The manufacturer not only shows social responsibility to the stakeholders and
shareholders, but also collects the used products from the third party and recycles them to new products. Considering
profit maximizing motives of the channel members, the paper examines the effect of manufacturer’s degree of social
responsibility on the collection activity of the third party. Under manufacturer Stackelberg game setting, it is found that
product recycling is directly proportional  to the manufacturer’s corporate social responsibility (CSR) concerns and there
must be a threshold of recycling for the optimal benefit that can be acquired through CSR practice. The proposed model
is illustrated by a numerical example and a sensitivity analysis reveals nature of the parameters.

Keywords  Closed-loop supply chain · Recycling · Corporate social responsibility · Game theory · Supply chain
coordination

1 Introduction the society, which, in turn, makes them a responsible

entity in a society. Companies are also bearing responsi-
In the modern corporate world, the concept of traditional bilities in making the society we live in, healthier where
business has undergone multiple changes. Production the so-called term ‘corporate Social Responsibility (CSR)’
houses have to take greater responsibility in addition to is used to entitle this set of activities. On the other hand,
making profit, and need to shoulder more responsibility the matter of scarce resources on the earth pushes com-
for the welfare of the society. For instance, many compa- panies to sustain some resources for the next generation.
nies are nowadays running hospitals for people, or extend- As for example, Tata Motors , the largest Indian automotive
ing their hands in community learning programs by estab- manufacturer, is trying to actively take part in CSR pro-
lishing educational institutes. Some other companies are grams providing support for the enhancement of schedule
conducting projects to uplift socially and economically caste and schedule tribe communities by Tata Affirmative
backward classes people. To do so, companies tend to Action Program (TAAP). Furthermore, Infosys supports non
share their profit with the members of supply chain and

*  Nikunja Mohan Modak, [email protected]; Sudipta Sinha, [email protected]; Shibaji Panda,

[email protected]; Nima Kazemi, [email protected] | 1Palpara Vidyamandir, Chakdaha, West Bengal 741222, India.
2
Department of Mathematics, Burdwan Raj College, University of Burdwan, Bardhaman, West Bengal 713104, India.
3
Department of Mathematics, Bengal Institute of Technology, 1. No. Govt. Colony, Kolkata 700150, India. 4Center for Transportation
and Logistics, Massachusetts Institute of Technology, Cambridge, MA 02142, USA.

SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

Received: 23 July 2019 / Accepted: 27 August 2019 / Published online: 10 September 2019

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

profitable organizations to eradicate hunger, poverty and profit in the centralized channel. Channel members in the
malnutrition. present model are contributing not only in CSR activities
Closed Loop Supply Chain (CLSC), a more environmen- but also in making an eco-friendly environment, and thus
tal friendly type of supply chain, has gained increasing information sharing among them will be more easier than
attention with the growing concern of protecting environ- those firms which have only profit maximizing motive.
ment and preserving natural resources. In fact, the impor- Thus, to establish a more efficient channel, we propose
tance of CLSC has been intensified by the pressure from a profit sharing mechanism in order to implement central-
environmental activist groups and regulations set up by ized decisions. Outcomes of the model are also investi-
authorities. CLSC includes all forward logistics flows (such gated through numerical examples and a comprehensive
as procurement of new raw materials, production and sensitivity analysis is presented to demonstrate effect of
distribution) along with the reverse logistics (collection parameters.
of used products and reuse). Recycling is a key element
of CLSC aims at turning the activities of a forward sup-
ply more sustainable, which indeed helps in maintaining 2 Literature review
ecological balance through eco-friendly activities. Recy-
cling is defined as the set of activities for breaking down or The alarming rate at which resources are currently being
reusing the consumed products (such as papers, glasses, exploited calls for reshaping the extraction and consump-
plastics, metals, batteries etc) that would otherwise be dis- tion of materials. Businesses are central part of the revo-
posed of as waste. However, collecting the used products lutionary movement toward a greener society, and there-
for the purpose of recycling from the waste has always fore they are trying to incorporate sustainable business
been a challenging activity in the reverse supply chain. It practices in all stages of their supply chain. This has led to
is evident that customers cannot return the used product creation of CLSC, which has been around for years, but it
due to the non-availability of the collection centers and did not attract much attention in the business world until
the lack of awareness of the evil effect of their disposal. the last decade since then companies started to realize the
They rather distort or waste the used items. So, establish- importance of their contribution in building more sustain-
ment of awareness campaigns among the end users is of able operations. CLSC is a revolutionized business model
crucial importance and the corporate houses as well as that has emerged through integrating the conventional
retailers have to come forward to make consumers aware supply chain flows with the reverse one, dealing with col-
of the eco-friendly effect of recycling. Therefore, introduc- lecting and processing End of Life (EOL) products. A very
tion/consideration of collectors of recyclable items in sup- important element of CLSC is traceability of elements
ply chains is very essential. Doing so not only benefits the flowing along the supply chain. Products, production
environment, but also helps companies to become more residues, part of the products must be identified in the
socially responsible as increasing the number of collec- original chain (tracking) and in the reverse flow (tracing).
tion centers can proportionally enhance the number of CLSC also seeks to maximize profit [13], through making
employments. a trade-off between the economic benefits of the supply
In the present paper, we develop a two-echelon CLSC chain, and its social and environmental effects.
comprising of a manufacturer, multiple retailers and a The literature on mathematical modeling of reverse
third party collector. The third party collects the used logistics and CLSC has advanced over the years since the
items from the customers and separates recyclable items seminal paper of Schrady [37], who developed an Eco-
through screening. He then makes an arrangement to nomic Order Quantity (EOQ) model with joint determi-
deliver the recyclable items to the manufacturer for fur- nation of production and recovery process. Since then,
ther use. The process of recycling thus lessens the crisis numerous numbers of studies have been published to ana-
of natural resources and maintains sustainability, which is lyze the structure of CLSCM and to examine it under differ-
the crying need of the hour for our successors. The paper ent business scenarios. For instance, through a case study
uses the concept of the Vicker’s [40] classic principle stat- and survey research, Cruz [6] studied an environmentally
ing that non-profit maximizing motives may earn higher responsible supply chain network using a multi-criteria
profits than the pure profit-maximizers. Here the objec- decision making approach. Yi [43] studied CLSC models
tives of the manufacturer is to engage in socially respon- under different market powers. Recycling and remanu-
sible activities and to explore its effects on recycling and facturing of waste products have become an important
profit making. To analyze different aspects of the manu- strategic operation of a modern enterprise, and it may
facturer’s contribution in socially responsible activities, bring produce significant economic and social benefits
we discuss the individual profit of the channel members [11]. In this regard, Zheng et al. [45] presented a compari-
separately in the decentralized channel as well as the joint son between a conventional and a reverse supply chain in

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

the Betrand Competition. Pan et al. [26] analyzed reverse designed so that it triggers the enthusiasm of members
logistics costs and benefits to explore the advantage of to attain the defined objective of the CLSC. The issue of
reverse logistics over the traditional supply chain. Ma [17] CLSC coordination has extensively been studied in the lit-
developed a model to investigate the strategy of optimal erature of supply chain management, aiming at improv-
recycling channels for manufacturers under the case of ing the performance of members individually and jointly.
multiple recycling channels. A challenge for manufacturers Among them, Mafakheri and Nasiri [18] considered post-
is to convince customers (regardless of whether they are consumption activities like reusing, refurbishing, remanu-
the manager of the store, or the end users), that it is worth facturing, and recycling in a dynamic setting and depicted
to return the used products to a collection point (gate that a higher environmental performance can be achieved
keeping point). Decision on the destination of the product through the coordination using the revenue sharing
depends on the value that can be recovered, and also on mechanism. Yoo et al. [44] examined the effectiveness of
the determination of reusable items [39]. Wei and Wei [42] wholesale price, buy-back and quantity discount contracts
suggested that a third-party logistics service as a collec- to coordinate a CLSC. Panda et al. [30] introduced two dif-
tion center can lessen the capital investment, differentiate ferent sequential bargaining procedures for benefit shar-
management risk, and advance service level. Ramos et al. ing among the channel members of a distribution channel.
[33] proved that a collection center in a reverse logistics Dutta et al. [9] showed that the buy-back offer along with
can reduce environmental and social impacts on the envi- revenue sharing mechanism can generate better channel
ronment. Zheng and Wu [46] made a deep study on CLSC performance in a CLSC. Considering demand and collec-
with the third party recycler based on consumer prefer- tion disruptions, a different type of all unit quantity dis-
ence. Aligned with the literature, in this paper a third party count contract is suggested by Wang and Wang [41] to
collector has been included so as it facilitates the collec- coordinate a closed-loop supply chain. Modak et al. [21]
tion and screening of reusable processes for the manufac- used profit sharing mechanism to coordinate a two-ech-
turer. Qualified items after screening are shipped to the elon CLSC having deopolies retailers. Saha et al. [34] pro-
manufacturer for further use. However, disqualified items posed a three-way discount mechanism to coordinate a
have been destroyed. In our model recycling is considered. closed-loop dual-channel supply chain. Modak et al. [20]
CSR intends to ensure that companies show their investigated different structures of CLSC and used the
responsibilities toward their stakeholders such as chan- alternative offer bargaining strategy for channel coordi-
nel members, consumers and society. CSR is nowadays a nation. An overview of contract mechanism applied to
key factor in consumer and client’s decisions to choose a coordinate CLSC in the literature of reverse supply chain
company to work with. In 2002, a survey depicted that, 94 can be found in the review of Govindan et al. [12], Aydin
per cent of companies believe that the implementation et al. [4] and Kazemi et al. [15].
of a CSR strategy can produce real business benefits [10]. This paper studies a scenario in a CLSC consisting of a
The reputable companies such as GAP, Adidas, Walmart, manufacturer, multiple retailers where a third party logis-
Nike are among the famous brands that have successfully tics is responsible for collecting the used items. The paper
implemented various forms of CSR in their supply chains focuses on used product recycling in a socially responsi-
[2]. As a result, a large number of principle organizations in ble distribution channel with the purpose to explore con-
different supply chain networks are taking CSR and using nection between social and environment responsibili-
it as a tool for profit enhancement. On the other hand, cus- ties. This study considers generalized channel structure
tomers keen to pay more price for CSR promised products with a third party who operates the reverse channel. In
[3, 5]. There have been a couple of studies addressing CSR real supply chains, it is difficult for the retailer to collect
issue in the literature. Cruz [7] developed a decision sup- the used products along with its retail sale. This is due to
port system for modeling and analyzing of a CSR supply several reasons, such as lack of space, types of the used
chain network. Cruz [8] studied the effect of CSR activities products (plastic, glass, metal etc.), and lack of manpower.
in the alleviation of risks in supply chain. Agan et al. [1] That is why, third party collection of used plastic, glass and
identified that CSR affects environmental performance of metal products is more common. The manufacturer not
a supplier in supply chain. In recent years, supply chain only shows social responsibility to the stakeholders and
operations and coordination under the consideration of shareholders, but also collects the used products from the
CSR activities become an essential issue for every business third party and recycles them to new products. Consider-
entities [14, 19, 22–24, 28, 29]. ing profit maximizing motives of the channel members,
Successful implementation of a CLSC depends heav- the paper examines the effect of manufacturer’s degree of
ily on construction of a well-performed coordination, social responsibility on the collection activity of the third
required to take into account the interest of different party. Furthermore, the optimal solutions are determined
entities of the supply chain. The coordination should be analytically for both centralized and decentralized models

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

and it is additionally examined how profit sharing mecha- Assumptions 2  Lead time is equal to zero, i.e., the prod-
nism enhances the channel’s performance. uct flows from the upstream to the downstream channel
without any delay when demand is placed. The manufac-
turer follows lot-for-lot policy. The system running cost of
3 Notations and assumptions each channel member is normalized to zero because in
such case the qualitative results will not be altered.
For model formulation we used the following notations.
pi Selling price of ith retailer (i = 1, 2, … n), a decision Assumptions 3  The third party collects the used prod-
variable. ucts at a√rate r, (0 < r < 1) whose functional form is given
by r = I∕𝛽  , where I is the investment of the retailer
r Recycling factor of the third party, a decision in collection activity and 𝛽 is a scaling parameter. This
variable. assumption is well established [36] in CLSC literature.
I Investment in collection activities of the third The total cost of collection of the used products is then
party. a function of the return rate (r) of the used products
∑n ∑n
C(r) Total cost to collect the used product. and is given by C(r) = I + c0 r i=1 Di = 𝛽 + c0 r i=1 Di  ,
w Wholesale price of the manufacturer. i.e., the sum of investment in collection and the total
cm Marginal cost of the manufacturer. number of used products collected from customers. For
c0 Amount paid by the third party to the customer non-negative profits of the third party it is assumed that
for returning a used product. (c1 + c3 )E(x) − (c0 + c2 + c3 ) > 0 and cm > c4 > c1 > c0.
c1 Amount paid by the manufacturer to the third
party for returning a recyclable used product.
c2 Inspection cost of a used product for the third 4 Mathematical modelling
party.
c3 Disposal cost a non-recyclable product for the Consider a scenario where n retailers source a single prod-
third party. uct from a manufacturer. The supply chain is a closed loop
c4  Value of each recyclable item for the type which is equipped with the third party collector and
manufacturer. recycling facilities. The third party collects the used prod-
x Probability of recyclable item. ucts from customers. After collecting the used products,
(1 − x) Probability of non recyclable item. the third party inspects them to differentiate recyclable
𝜋ri Profit function of the ith the retailer (i = 1, 2, … n). and non-recyclable items. It then sends recyclable items
𝜋t Profit function of the third party. to the manufacturer and disposes non-recyclable items.
𝜋m Pure profit function of the manufacturer. Channel structure of the proposed model is presented in
vm Total profit function of the manufacturer. Fig. 1.
pc Retail price of the product in centralized decision Many leading brands face intense pressure to involve
𝜋c Pure profit function of centralized channel in socially responsible activities [2]. A commonly noted
vc Total profit function of centralized channel response to this pressure is that the primary firm intro-
E(.) Expectation of ‘.’ duces code of conduct to its business partners that moti-
vates or forces them to be more socially responsible [32].
Thus, we assume that the manufacturer invests in CSR and
Assumption  Before the analytical discussion of the math- aligns its CSR goal with the channel performance. The cost
ematical model, we state our key assumptions that are par- associated with the CSR is shared among all the channel
ticularly defined for the model. members through a transfer pricing [6]. In modeling and
analysis, we only consider effects of CSR in the form of
Assumptions 1  The manufacturer is socially responsi- consumer surplus rather than the CSR activities. It is well
ble as well as Stackelberg leader of the channel. Retail- established that a firm’s CSR is accounted through con-
ers are non-competitive. Demand of the product at ith sumer surplus of it’s stakeholders [16, 25]. Consumer sur-
retailer’s end is assumed to be linearly price sensitive, plus is the difference between the maximum price that
that is, Di = ai − bi pi . Where, ai > 0 is the market poten- the consumers are willing to pay for a product and the
tial and bi > 0 is the customer’s price sensitivity factor. market price that they actually pay for the product. Thus,
pi ∈ (0, ai ∕bi ) ensures that under any circumstance the the consumer surplus is
demand at the ith retailer’s end is non-negative. Linear [ ]
price dependent demand is common and well-established ∑n pi∕max ∑
n

∫
Di (pi )dpi = [(Di2 ∕2bi )] (1)
in supply chain literature [27, 31]. i=1 pi∕mtk i=1

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

Fig. 1  Channel structure of the model under study

where pi∕max and pi∕mtk respectively denotes the maximum pure profit maximizer and 𝜃 = 1represents that the manu-
price that the consumers are willing to pay for a product facturer is the perfect welfare maximizer.
and the market price that they actually pay for the prod-
uct at ith retailer’s end. Suppose, 𝜃 ∈ [0, 1] is the fraction 4.1 Decentralized decision
of CSR that the socially responsible manufacturer is con-
∑n
cerned, then it incorporates 𝜃 i=1 [(Di2 ∕2bi )] as consumer In decentralized decision making, the channel members
surplus in its profit. 𝜃 = 0 implies that the channel is the operate independently and maximize their individual
profit functions. The manufacturer is the leader of the

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

channel and other members are its follower. The manufac- Using backward induction, the Stackelberg equilibrium
turer initiates the first move and enforces its own strategy solution in the decentralized decision making process
on the third party and retailers. Based on the manufactur- can be found as follows. First, each of the n retailers
er’s strategy, the third party and retailers find their move will announce their individual selling price to maximize
that maximizes their own profit. In decentralized decision their respective profits for a given wholesale price, w
making context, profit function of the ith retailer is of the manufacturer. Solving the necessary condition
d𝜋ri (pi )∕dpi = 0 , to optimize the profit function of the ith
𝜋ri (pi ) = (pi − w)Di (2) retailer (i = 1, 2, … , n) yields
Profit function of the third party is
ai + bi w
∑
n
∑
n pi = , i = 1, 2, 3, ..n (9)
𝜋t = xc1 r Di − (1 − x)c3 r Di 2bi
i=1 i=1
(3) Note that, d 2 𝜋ri ∕dpi2 = −2bi < 0 , for all i = 1, 2, , n . That is,
∑n
all 𝜋ri are concave function of pi i = 1, 2, … , n . Following n
− (c0 + c2 )r Di − 𝛽r 2
i=1 retailers reaction, the third party will optimize its expected
profit function and will determine the optimum level of
recycling facility. Now solving the necessary condition
Pure profit function of the manufacturer is given by
dE(𝜋t )∕dr = 0 , we get
∑
n
∑
n
∑n ∑n
𝜋m = (w − cm ) Di + (c4 − c1 )xr Di (4) ( i=1 ai − ( i=1 bi )w)((c1 + c3 )E(x) − (c0 + c2 + c3 ))
i=1 i=1 r=
4𝛽
(10)
As a result, the total profit function of the manufacturer
Moreover, note that, d E(𝜋t )∕dr = −𝛽 < 0 that is, E(𝜋t ) is a
2 2
is calculated as
concave function of r and so the value of r given in Eq. (10)
∑
n
∑
n
will maximize profit function of the third party. Following
vm (w) = (w − cm ) Di + (c4 − c1 )xr Di decisions of the n retailers and the third party, finally the
i=1 i=1
(5) manufacturer makes its decision. Since the manufacturer
∑
n
is involved in CSR, there exist two parts (pure profit and
+𝜃 [(Di2 ∕2bi )]
i=1 consumer surplus) in the manufacturer’s total expected
profit. As a result. the socially responsible manufacturer
The expected profits of the third party and the expected will maximize its expected total profit and the necessary
pure and the total profit of the manufacturer are condition of optimization i.e., dE(vm )∕dw = 0 yields
∑n ∑n ∑n
( i=1 ai (𝛽(2 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))) + 2 i=1 bi cm 𝛽)
w∗ = ∑n ∑n (11)
b (𝛽(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )))
i=1 i

∑n
b
∑
n
∑
n Moreover, note that d 2 E(vm )∕dw 2 = − i=1 i
(𝛽(4 − 𝜃)
∑n 4𝛽
E(𝜋t ) = E(x)c1 r Di − (1 − E(x))c3 r Di − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))) . Thus
i=1 i=1
(6) the expected total profit of the manufacturer ( E (vm ) ) will
∑
n ∑n
2 be a concave function of w if (𝛽(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)
− (c0 + c2 )r Di − 𝛽r
i=1
((c1 + c3 )E(x) − (c0 + c2 + c3 ))) > 0 . From the above discus-
sion we have the following proposition.

∑
n
∑
n Proposition 1  The total expected profit function of
E(𝜋m ) =(w − cm ) Di + (c4 − c1 )E(x)r Di (7) the manufacturer (E(vm )) is a concave function of w if
∑n
i=1 i=1 (𝛽(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))) > 0.

∑
n
∑
n Now, by inserting the optimal value of the manufactur-
E(vm ) = (w − cm ) Di + (c4 − c1 )E(x)r Di er’s wholesale price into (10) and (9), we have the optimal
i=1 i=1
(8) values of recycling rate and selling price as follows
∑
n
+𝜃 [(Di2 ∕2bi )]
i=1

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

� ∑n ∑n �
( i=1 ai − i=1 bi cm )((c1 + c3 )E(x) − (c0 + c2 + c3 ))
∗
r = min ∑n ,1 (12)
2(𝛽(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )))

ai
pi∗ =
2bi
∑n ∑n ∑n (13)
( i=1 ai (𝛽(2 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))) + 2 i=1 bi cm 𝛽)
+ ∑n ∑n , i = 1, 2, 3, ..n
2 i=1 bi (𝛽(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )))

∑n
∑n ∑n
Note that, dr ∗ ∕dE(x) = ( ai − i=1 bi cm )( i=1 bi (c4 − c1 )
i=1 4.2 Centralized decision
((c1 + c3 )E(x) − (c0 + c2 + c3 ))2 + (c1 + c3 )𝛽(4 − 𝜃))∕(2(2(𝛽(4 − 𝜃)
∑n
− i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))))2 ) > 0.  It In centralized decision making, the channel members
indicates that, the optimal recycling rate of the third party operate jointly, i.e., supply chain makes integrated deci-
increases with the increase of the expected value of the sion. A single decision maker took all decisions to maxi-
recyclable items. The result is straightforward because the mize the total channel profit. Under the centralized sce-
higher value of the expected number of recyclable items nario, the expected pure profit and the total expected
from the collected used products provides more profit to profit of the channel are respectively given by
the third party, which an incentive that motivates it to col-
lect more used items. Substituting optimal values from ∑
n
∑
n
E(𝜋c ) = (pci − cm )Di + (c4 − c1 )E(x)r Di − 𝛽r 2 (16)
(11)–(13) in (2), (6)–(8) we get optimal profits of the chan-
i=1 i=1
nel members.
From Eq.  (12), it can be concluded that if
∑n ∑n ∑
n
∑
n
( i=1 ai − i=1 bi cm )((c1 + c3 )E(x) − (c0 + c2 + c3 )) > 2(𝛽 E(vc ) = (pci − cm )Di + (c4 − c1 )E(x)r Di
∑n
(4 − 𝜃) − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 ))) then i=1 i=1
r ∗ = 1 , entailing that the third party collects 100% of the (17)
∑
n
used products for recycling. In this case, the modified +𝜃 [(Di2 ∕2bi )] − 𝛽r 2

expected profit of the third party, expected pure and the i=1

expected total profit of the manufacturer can be found

from Eq. (6)–(8) respectively by just replacing r = 1 . In this Solving the necessary conditions to optimize the total
special situation, the optimal wholesale and selling prices expected profit function of the channel i.e., solving
of the decentralized channel are as follows 𝜕E(vc )∕𝜕pci = 0 i = 1, 2, … , n and 𝜕E(vc )∕𝜕r = 0 gives opti-
∑n ∑n ∑n mal selling prices and recycling rate as follows.
∗
(2 − 𝜃) i=1 ai + 2 i=1 bi cm −2 i=1 bi (c4 − c1 )E(x)
wsp = ∑n pci∗ =
(1 − 𝜃)ai + bi cm
b (4
i=1 i
− 𝜃) bi (2 − 𝜃)
(14) ∑n ∑n
bi ( i=1 ai − c i=1 bi )(c0 + c2 + c3 − (c3 + c4 )E(x))2
− � ∑n �
ai bi (2 − 𝜃) 2𝛽(2 − 𝜃) − i=1 bi (c0 + c2 + c3 − (c3 + c4 )E(x))2
∗
pi∕sp =
2bi (18)
∑n ∑n ∑n
(2 − 𝜃) i=1 ai + 2 i=1 bi cm − 2 i=1 bi (c4 − c1 )E(x) � ∑n ∑n �
+ ∑n , ( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 )
2 i=1 bi (4 − 𝜃) rc∗ = min ,1
2𝛽(2 − 𝜃) − B(c0 + c2 + c3 − (c3 + c4 )E(x))2
i = 1, 2, , n (19)
(15)
The optimal selling prices ( pci∗  ) and recycling rate ( rc∗ ) will
Hence, substituting pi = pi∕sp
∗
 , w = wsp
∗ and r = 1 in (2), maximize the total expected profit function of the channel
(6)–(8), we get the optimal expected pure and total profit if E(vc ) satisfies concavity conditions. The following propo-
of the integrated channel in this special case. sition depicts concavity conditions of E(vc ).

Proposition 2  The total expected profit func-

tion of the channel (E(vc )) is a concave func-
tion of the decision variables ( p1  , p2  , ..., pnand r) if

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1
∑n
2b1 b2 … bn 𝛽(2 − 𝜃)n − b1 b2 … bn T 2 (2 − 𝜃)(n−1) ( i=1 bi ) > 0 the third party will collect more used products for recy-
where n = 1, 2, … denotes number of retail outlets. cling. As a consequence, the recycling factor attains the
minimum value when the manufacturer is purely profit
Proof  Please see “Appendix 1”. maximizer i.e., 𝜃 = 0 and it attains maximum value when
the manufacturer is the perfect welfare maximizer, i.e., 𝜃 =
By substituting the optimal values from (18), and (19) 1, for which the recycling factor may be higher than one.
in (16) and (17), we obtain the optimal expected pure This is quite obvious because the retailer has no obliga-
and the total profit of the integrated channel. Note that, tion in participating in the reverse supply chain through
∑n ∑n
if ( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 ) > [2𝛽 collection of the used products when the manufacturer
∑n
(2 − 𝜃) − i=1 bi (c0 + c2 + c3 − (c3 + c4 )E(x))2 ]then rc∗ = 1 . is not taking up CSR. Nevertheless, the recycling factor is
In such a situation modified expected pure profit and total bounded by 0 and 1. Thus, the maximum and the mini-
expected profit of the channel can be found from Eqs. (14) mum value of the recycling factor in decentralized sce-
and (15) respectively by replacing r = 1 . In this case, the nario are respectively as follows.
� ∑n ∑n �
( i=1 ai − i=1 bi cm )((c1 + c3 )E(x) − (c0 + c2 + c3 ))
∗
rdmin = min ∑n ,1 (22)
2(4𝛽 − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )))

� ∑n ∑n �
( i=1 ai − i=1 bi cm )((c1 + c3 )E(x) − (c0 + c2 + c3 ))
∗
rdmax = min ∑n ,1 (23)
2(3𝛽 − i=1 bi (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )))

On the other hand, differentiating the centralized optimal

optimal selling price of the centralized channel at ith retail-
recycling rate with respect to 𝜃 , we get
er’s end is given by
∑n ∑n
drc∗ 2𝛽( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 )
a (1 − 𝜃) + bi cm − b1 ((c3 + c4 )E(x) − c0 − c2 − c3 ) = >0
d𝜃 [2𝛽(2 − 𝜃) − B(c0 + c2 + c3 − (c3 + c4 )E(x))2 ]2
∗
pci∕sp = i ,
b1 (2 − 𝜃) (24)
i = 1, 2, , n From Eq. (24) it can be understood that the optimal recy-
(20)
cling rate in the centralized scenario also behaves in simi-
Hence, substituting pci = pci∕sp
∗
and r = 1in (16) and (17) we lar manner as in the decentralized scenario with respect
drive the optimal expected pure and total profit of the to 𝜃 . The maximum and the minimum value of the recy-
integrated channel in this special case. We will now discuss cling factor in the decentralized scenario are respectively
the effect of the degree of social concerns of the manufac- as follows.
turer on the optimal decisions. ∗
rcmin
� ∑n ∑n �
4.3 Analyzing the effect of the degree of CSR = min
( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 )
,1
4𝛽 − B(c0 + c2 + c3 − (c3 + c4 )E(x))2
As discussed before, the manufacturer acts as a Stack- (25)
elberg leader of the channel and shows CSR. The manu-
facturer can enhance its CSR program by highlighting ∗
rcmax
� ∑n ∑n �
recycling activity. Thus, there exists an intra-relationship ( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 )
between the recycling rate and degree of the manufac- = min ,1
2𝛽 − B(c0 + c2 + c3 − (c3 + c4 )E(x))2
turer’s CSR activity. Note that, (26)
∑n ∑n
dr ∗ ( i=1 ai − i=1 bi cm )((c1 + c3 )E(x) − (c0 + c2 + c3 ))𝛽
= ∑n (21)
d𝜃 2(𝛽(4 − 𝜃) − (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )) i=1 bi )2

∗
Clearly, dr
d𝜃
> 0 that is optimal recycling rate increases with
the increasing degree of CSR activity of the manufacturer.
If the manufacturer is more concerned about the CSR, then

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

A graphical representation of r ∗ and rc∗ with respect to 𝜃 is

demonstrated in Fig. 2. Numerical value of the parameters
are considered the same as the numerical section.
From the above discussion, we can derive the following
proposition.

Proposition 3  In both centralized and decentralized sce-

narios the optimal recycling rate increases with the increas-
ing concerns of the CSR activity and it lies in the range ( rdmin∗
 ,
rdmax  ) and ( rcmin , rcmax  ) respectively for the centralized and
∗ ∗ ∗

decentralized scenarios.

Figure 2 reveals that, the decentralized recycling rate

increases with the growth in the value of 𝜃 . On the other
hand, recycling rate in the centralized model is equal to 1 for
all values of 𝜃 ∈ [0, 1] because rcmin
∗
= min {2.91137, 1} = 1
and rcmax = min {17.1588, 1} = 1 . Moreover, as dw ∗ ∕d𝜃 < 0
∗
Fig. 3  Effect of 𝜃 on profits
and pi∗ ∕d𝜃 < 0 (please see “Appendix 2”), the wholesale
price of the manufacturer decreases if value of 𝜃 elevates.
These results show that increment in the consumer surplus
It leads us to conclude that the selling price decreases but
overestimates the loss in the pure profit.
demand increases with the increasing value of 𝜃 . Figure 3
demonstrates the effect of 𝜃 on individual profits, as well
as the pure and the total profit of the manufacturer.
From Fig. 3 it is clear that retailers and the third party
5 Cooperative solution and profit sharing
obtain more profits for higher value of 𝜃 . They will receive
mechanism
the minimum profits when 𝜃 = 0 and the maximum prof-
Cooperative integrated decision is more profitable than
its for 𝜃 = 1 . Thus, purely profit maximizing motive of
the non-cooperative decision due to non occurrence of
the manufacturer provides the minimum profit to the
the well known ‘double marginalization’ problem. This sec-
other channel members while totally welfare maximizing
tion demonstrates an approach to adopt the centralized
motive gives them the opportunity to realize the maxi-
decision through a profit sharing mechanism. For accept-
mum profits. Furthermore, the pure profit of the manu-
ing the centralized order quantity, which is more than the
facturer decreases with the increasing value of 𝜃 , yet the
decentralized one, the retailer’s cost will increase and there
total profit of the manufacturer increases with respect to 𝜃 .
is no reason that the retailer will adopt the centralized
policy unless it has proper incentive. As an incentive, the
manufacturer can offer the retailers and the third party to
share the surplus profit if they adopt the centralized deci-
sions. Under the profit sharing mechanisms, the system
performance is first optimized and the resultant benefit is
then shared among the channel members. Note that, the
total expected profit in the decentralized scenario is equal
∑n
to E(vm )∗ + E(𝜋t )∗ + i=1 𝜋ri∗ . Thus, if the channel members
adopt the centralized decision then the channel generates
∑n
surplus profit, Δsp = E(vc )∗ − [E(vm )∗ + E(𝜋t )∗ + i=1 𝜋ri∗ ] .
Under the proposed profit sharing mechanism, the manu-
facturer offers the channel members to share surplus profit
according to their decentralized profit if they adopt cen-
tralized decisions. Hence, under the proposed contract,
profit of the ith retailer, the third party and the manufac-
turer can be found as follows

𝜋ri∗ Δsp
𝜋risp = 𝜋ri∗ + ∑n i = 1, 2, , n (27)
Fig. 2  Effect of 𝜃 on optimal recycling rates E(vm )∗ + E(𝜋t )∗ + i=1
𝜋ri∗

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

Table 1  The optimal values in Optimal Decentralized channel Centralized channel

decentralized and centralized
model R1 R2 R3 R4 Third party Manufacturer

Wholesale price – – – – – 139.5 –

Selling price 169.7 174.7 171.7 170.7 – – 122.7, 126.5, 124.2, 123.5
Recycling rate – – – – 0.6727 – 1
Demand 15.1 17.6 16.1 15.6 – – 38.6, 41.8, 39.9, 39.2
Profit 457.6 621.4 520.1 488.4 316.8 3355.6 9484.8

E(𝜋t )∗ Δsp
𝜋tsp = E(𝜋t )∗ + ∑n (28)
E(vm )∗ + E(𝜋t )∗ + i=1
𝜋ri∗

sp
E(vm )∗ Δsp
𝜋m = E(vm )∗ + ∑n (29)
E(vm )∗ + E(𝜋t )∗ + i=1
𝜋ri∗
sp ∑n
Note that, 𝜋m + 𝜋tsp + i=1 𝜋risp = E(vc )∗ . That is, the profit
sharing scheme not only provides the best channel out-
comes, but also assures the win–win profit for all the chan-
nel members. The next section illustrates a numerical
example of the proposed model, followed by a sensitivity
analysis.

Fig. 4  Decentralized versus coordinated profit for different 𝜃

6 Numerical illustration

Consider a distribution channel consisting of a manufac- to 5759.91, which is 64% less than the centralized chan-
turer (M), and four retailers ( R1 , R2 , R3 , andR4 ). The param- nel’s expected profit. Thus, channel coordination is impera-
eter values are considered as a1 = 100 , a2 = 105 , a3 = 102 , tive to improve its performance. Under the proposed profit
a4 = 101 , b1 = b2 = b3 = b4 = 0.5 , c0 = 10 , c1 = 30 , c2 = 1 , sharing mechanism channel members have the opportu-
c3 = 14 , c4 = 40 , cm = 100 , 𝛽 = 700 𝜃 = 0.4 , Suppose the nity to improve the channel’s profit equal to 3724.9. Profit
random variable x follows the uniform distribution over of the four retailers will be 753.53, 803.71, 672.69, and
the interval [0.80, 1.0], then E(x) = 0.90 . Based on the pre- 804.24 respectively if they adopt cooperative decisions.
sent parameter setting the optimal solutions are calcu- Profit of the third party and the manufacturer under this
lated and presented in Table 1. mechanism are 521.67 and 5525.65 respectively.
The optimal solutions of the centralized channel displayed Figure 4 depicts a comparison of individual profit of the
in the Table 1 are determined from the special case because channel members between decentralized and coordinated
∑n ∑n
( i=1 ai − cm i=1 bi )((c3 + c4 )E(x) − c0 − c2 − c3 ) = 4908.8 situation (in Fig. 4, without loss of generality we have con-
∑n
> [2𝛽(2 − 𝜃) − i=1
bi (c0 + c2 + c3 − (c3 + c4 )E(x))2 ] = 1126.08  . sidered one retailer instead of all retailers because all the
Owing to the fact that rc∗ = 1 > r ∗ = 0.6727 , the central- retailers have similar output in comparison). Clearly, profit
ized channel is more environmental friendly than the sharing mechanism provides win–win opportunity for all
decentralized supply chain. Consumers also prefer the cen- the channel members. Next subsection provides a sensitiv-
tralized model as they have to pay less as a consequence ity analysis on the model parameters.
demand in the centralized channel is much higher than
the decentralized channel. The expected pure profit and 6.1 Sensitivity analysis
consumer surplus of the manufacturer in the decentralized
scenario is equal to 2938.16 and 417.44. While in the cen- Sensitivity analysis of the parameters is presented in
tralized scenario, the expected pure profit and consumer Table 2. From Table 2, it can be observed that the market
surplus of the channel is equal to 6938.6 and 2546.2. The potentiality and price sensitivity parameters have high
expected total profit in the decentralized channel is equal effect on the profits of the channel members. The param-
eter c0 highly affects recycling rate and as a result it has

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

Table 2  Sensitivity of the key parameters

Changes (%) w r p D 𝜋ri E(𝜋t ) E(𝜋m ) E(vm ) pc Dc E(𝜋c ) E(vc )

ai −  40 −  5.45 −  19.4 −  25.81 −  119.2 −  96.13 −  34.76 −  34.76 −  30.69 −  24.37 −  14.76 −  22.59 −  22.04
−  20 −  2.72 −  8.95 −  12.9 −  59.6 −  83.87 −  42.18 −  18.31 −  17.16 −  12.14 −  32.38 −  13.99 −  15.81
+ 20 + 2.72 10.45 12.9 60.26 155.47 16.79 20.16 20.8 12.3 32.38 19.4 18.93
+ 40 5.45 19.4 25.81 119.87 227.07 42.18 42.16 45.3 24.53 64.77 44.21 43.12
bi −  40 6.81 8.96 42.13 82.78 456.86 17.43 32.18 38.05 40.83 24.87 41.10 40.08
−  20 3.15 4.48 16.09 40.4 145.61 8.59 15.05 0.98 141.5 12.44 14.92 14.55
20 −  2.95 −  4.48 − 10.96 −  37.75 −  68.09 −  8.3 −  13.27 −  12.24 −  10.19 −  12.44 −  9.28 −  9.05
40 −  5.52 −  7.46 −  19.09 −  74.17 −  95.34 −  16.36 −  24.98 −  22.08 17.44 −  24.61 −  15.96 −  14.98
c0 −  40 −  1.43 31.34 −  0.59 3.31 7.14 73.19 2.81 3.3 −  2.04 3.37 7.0 6.83
−  20 −  0.79 16.42 −  0.29 1.99 3.5 33.53 1.38 1.63 −  0.98 1.56 3.48 3.39
20 0.72 −  14.92 0.29 −  1.32 −  3.32 −  27.82 −  1.35 −  1.57 1.06 −  1.56 −  3.42 −  3.34
40 1.43 −  29.85 0.59 −  3.31 −  6.49 −  50.49 −  2.67 −  3.09 2.04 −  3.11 −  6.69 −  6.62
c1 −  40 2.14 −  303.33 0.88 −  4.97 −  10.20 −  1526.74 −  4.28 −  4.97 – – – −
−  20 −  0.04 −  58.54 −  0.02 0.10 0.19 −  151.3 0.08 0.10 – – – –
20 2.26 23.16 0.93 −  5.26 −  10.6 40.95 −  4.52 −  5.26 – – – –
30 3.97 29.36 1.65 − 9.83 −  20.63 50.1 −  8.45 −  9.83 – – – –
c2 − 40 −  0.14 2.98 −  0.06 0.66 0.68 6.25 0.27 0.32 −  0.16 0.52 0.69 0.67
−  20 −  0.07 1.49 0 0.66 0.35 3.09 0.14 0.16 −  0.08 0.26 0.34 0.34
20 0.07 −  1.49 0.06 0 −  0.33 −  2.99 −  0.14 −  0.16 0.08 0 −  0.34 −  0.34
40 0.14 −  2.98 0.06 0 −  0.68 −  5.97 −  0.28 −  0.32 0.24 −  0.26 −  0.69 −  0.67
c3 −  40 −  0.22 4.48 −  0.06 0.46 0.94 8.81 0.38 0.45 −  0.24 0.51 0.97 0.94
−  20 −  0.07 2.98 0 0.26 0.48 4.36 0.19 0.23 −  0.08 0.26 0.48 0.47
20 0.71 −  1.49 0.06 −  0.26 −  0.48 − 4.2 −  0.19 −  0.22 0.16 −  0.26 −  0.48 −  0.47
40 0.21 −  4.47 −  0.12 −  0.53 −  0.94 −  8.33 −  0.38 −  0.44 0.32 − 0.52 −  0.96 − 0.94
c4 −  20 3.94 − 7.46 1.65 −  8.6 − 17.35 −  16.29 − 7.42 −  8.52 3.67 − 5.7 −  12.06 − 11.77
−  10 2.01 − 4.66 0.82 − 4.66 − 9.53 −  9.53 − 4.00 −  4.65 1.87 − 3.17 −  6.86 − 6.37
20 −  4.73 10.45 − 1.94 11.26 23.1 21.63 8.66 10.27 −  3.67 5.96 12.76 12.45
40 −  10.61 23.88 − 4.3 24.5 54.76 51.06 18.91 22.88 −  7.25 11.66 26.23 25.58
cm −  40 − 17.78 38.81 −  7.31 41.06 98.82 91.76 91.71 91.68 −  29 32.38 79.77 77.81
−  20 − 8.89 19.4 −  3.65 20.53 45.21 42.18 42.16 42.14 −  10.19 16.32 37.18 36.27
20 8.89 −  19.4 3.65 − 20.4 −  36.8 − 34.76 −  34.76 − 34.75 10.19 −  16.06 − 31.78 −  30.97
40 17.78 −  38.8 7.36 − 41.06 −  65.19 − 62.14 −  62.13 − 62.10 20.46 −  32.38 −  58.15 −  56.72

manufacturer are less sensitive on c0 . We have set up 30%

increment of c1 instead of 40%, and 10%, 20% decrement
of c4 because of the assumption c4 > c1 . c1 is not involved
in the centralized policy and so it has no impact on the
centralized decisions.
The sensitivity of the parameter c1 is very interest-
ing. If we consider c1 , the amount paid by the manufac-
turer to the third party for returning a recyclable used
product as a decision variable then the total expected
profit of the manufacturer will be maximum when it sets
c1 = [c0 + c2 + (1 − E(x))c3 + c4 E(x)]∕2E(x) . In the present
illustration the total expected profit of the manufacturer
will be maximum if c1 = 26.89 (see Fig. 5). Wholesale and
Fig. 5  Nature of the total expected profit of the manufacturer with selling prices increase (or decrease) moderately with the
respect to c1
increment (or decrement) of cm but recycling rate behaves
in an opposite manner. Both c2 and c3 have little impact
a high impact on the expected profit of the third party. on profit of the channel members except the third party.
However, the wholesale price, the selling price, profit of Total expected profit of the manufacturer in decentralized
retailers, and the expected pure and the total profits of the

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

channel and total expected profit of the centralized chan- The present work also reveals following insights. The
nel are highly affected by the parameter c4. value of each recyclable item for the manufacturer has
noticeable impact on the total expected profit of the
manufacturer in decentralized channel as well as total
7 Concluding remarks expected profit of the centralized channel. That is, value
of the material extracted from used products is a key factor
In this paper, we developed a socially responsible CLSC to operate the reverse channel supply chain. The amount
including a manufacturer, multiple retailers and a third paid by the manufacturer to the third party for returning
party collector. The third party entity is responsible for col- a recyclable used product ( c1 ) is required to determine
lecting the used products from customers. The collector carefully to maximize the channel leader’s expected profit
carries out inspection on the used products to differenti- and the present model provides value of c1 in closed form.
ate recyclable and non-recyclable items, and then sends Market potential and marginal cost of the manufacturer
recyclable items to the manufacturer and disposes non- have high impact on the total expected profits. The manu-
recyclable items. The developed model can be taken as facturer’s social welfare activity motivates the used prod-
a simulator of the supply chain of some specific products uct collection activity of the third party. Manufacturer’s
like plastic, glass and metals. In the proposed model, the awareness activities to enhance recycling of used products
manufacturer acts as a Stackelberg leader, who is willing can influence customers to return used products through
to take part in corporate social responsibility schemes. third party. Awareness program can be consider as a part
The type of CSR which is taken in this study in the form of of CSR activity. The work also reveals that inefficiencies in
consumer surplus rather than any CSR activities. Hereby, decentralized channel can be possible to overcome using
the manufacturer has two motives which are pure profit the profit sharing mechanism.
and welfare maximization. To investigate the effect of the Although the proposed model provides some impor-
CSR activity of the manufacture, we analyzed the model tant insights to the field, it is not free of limitations. First,
to explore how manufacturer can set up its price and recy- demand is assumed to be deterministic and dependent
cling rate to ensure maximizing the profit for both decen- on price. Due to the variation of demand in real cases, it
tralized and centralized channels. It is observed that the will be more realistic if the demand is assumed to be the
expected total profit in the centralized system is much probabilistic. Second, we assumed that there are n retailers
more than the one in the decentralized channel. It is fur- in the channel who are non-competitive. As the retailers
ther concluded that the optimal recycling rate increases may be in competition in some scenarios, thus, the model
with increasing the CSR activity of the manufacturer. More- can be extended to the case of competitive retailers. Third,
over, we found that the profit sharing mechanism provides this model can be extended in view of carbon emission
best channel performance, giving a win–win profit to all reduction [35] and solid waste management [38]. Fourth,
the channel members. Finally, the numerical example the model can be examined under different game theo-
showed that expected total profit function of the manu- retical aspects like evolutionary game, sub-game perfect
facturer is a concave function of c1 and the manufacturer equilibrium and strategic bargaining. Moreover, some
may treat it as decision variable if the third party agrees. other scenarios such as: manufacturer disruptions due
The paper at hand makes several contribution to the to reliability of manufacturing system, imperfect quality
field, some of which are as follows. First, the paper is devel- items, proper cleaning and disinfections of the recycled
oped under a realistic closed-loop channel structure with items can be investigated in future research.
multiple retailers and a third party collector. Second, the
manufacturer shows social responsibility and non-profit
maximization motive (i.e., social welfare motive), which Compliance with ethical standards 
are reflected through consumer surplus. Third, the model
examines the effect of the degree of CSR on the decen- Conflict of interest  There is no conflict of interest in this article.
tralized and centralized channel structures. Fourth, the
optimal selling prices in each retailer’s end and recycling
rate are determined analytically taking into account the Appendix 1
restriction of recycling rate. Higher concerns about the
CSR ensures more collection of used products. Finally, the To check the concavity of total expected profit function
optimal ranges of recycling rate are determined for both of the channel, we compute its corresponding Hessian
decentralized and centralized models. matrix, Hc as follows

Vol:.(1234567890)
SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1 Research Article

�∑n ∑n �
dpi∗ ai − bi cm 𝛽 2
⎛ 𝜕 2 E(vc ) 𝜕 2 E(vc ) 𝜕 2 E(vc )
− −
𝜕 2 E(vc ) ⎞ =− i=1 i=1
∑ <0
d𝜃 (𝛽(4 − 𝜃) − (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )) ni=1 bi )2
⎜ 𝜕p12 𝜕p1 𝜕p2 𝜕p1 𝜕p3 𝜕p1 𝜕r ⎟ �∑n ∑n � 2
⎜ 𝜕 2 E(vc ) 𝜕 2 E(vc ) 𝜕 2 E(vc )
− −
𝜕 2 E(vc ) ⎟ dw ∗ 2 i=1 ai − i=1 bi cm 𝛽
⎜ 𝜕p2 𝜕p1 𝜕p22 𝜕p2 𝜕p3 𝜕p2 𝜕r ⎟ d𝜃
=− ∑
(𝛽(4 − 𝜃) − (c4 − c1 )E(x)((c1 + c3 )E(x) − (c0 + c2 + c3 )) ni=1 bi )2
<0
⎜ 𝜕 2 E(vc ) 𝜕 2 E(vc ) 𝜕 2 E(vc )
− −
𝜕 2 E(vc ) ⎟
Hc = ⎜ 𝜕p3 𝜕p1 𝜕p3 𝜕p2 𝜕p32 𝜕p3 𝜕r ⎟
⎜ − − − − − − ⎟
⎜ ⎟ References
⎜ − − − − − − ⎟
𝜕 2 E(vc ) 𝜕 2 E(vc ) 𝜕 2 E(vc ) 𝜕 2 E(vc )
⎜ − − ⎟ 1. Agan Y, Kuzey C, Acar MF, Aikgz A (2016) The relationships
⎝ 𝜕r𝜕p1 𝜕r𝜕p2 𝜕r𝜕p3 𝜕r 2 ⎠
between corporate social responsibility, environmental sup-
plier development, and firm performance. J Clean Product
112:1872–1881
That is, 2. Amaeshi KM, Osuji OK, Nnodim P (2008) Corporate social
responsibility in supply chains of global brands: a boundary
⎛ −b1 (2 − 𝜃) 0 0 − − b1 T ⎞ less responsibility? Clarifications, exceptions and implications.
⎜ ⎟
⎜ 0 − b2 (2 − 𝜃) 0 − − b2 T ⎟ J Bus Ethics 81(1):223–234
⎜ 0 0 − b3 (2 − 𝜃) − − b3 T ⎟⎟ 3. Auger P, Burke P, Devinney TM, Louviere JJ (2003) What will con-
Hc = ⎜ sumers pay for social product features? J Bus Ethics 42:281–304
⎜ − − − − − − ⎟
⎜ ⎟ 4. Aydin R, Kwong CK, Ji P (2016) Coordination of the closed-loop
⎜ − − − − − − ⎟ supply chain for product line design with consideration of
⎜ ⎟ remanufactured products. J Clean Prod 114:286–298
⎝ b1 T b2 T b3 T − − − 2𝛽 ⎠
5. Cotte J, Trudel R (2009) Socially conscious consumerism: a sys-
where T = (c0 + c2 + c3 − (c3 + c4 )E(x)) . If the principal tematic review of the body of knowledge. Network for business
minors are alternatively negative and positive, i.e., the kth sustainability knowledge project series
6. Cruz JM (2008) Dynamics of supply chain networks with cor-
order leading principal minor Δk follows the sign of (−1)k , porate social responsibility through integrated environmental
then the total expected profit function of the channel decision-making. Eur J Oper Res 184(3):1005–1031
( E (vc ) ) will be concave, i.e., maximum at ( pci∗  , rc∗). 7. Cruz JM (2009) The impact of corporate social responsibility
| | in supply chain management: multicriteria decision-making
| −b1 (2 − 𝜃) 0 | approach. Decis Support Syst 48(1):224–236
Here Δ1 = 𝜕 2 E(vc )∕𝜕p12 = −b1 (2 − 𝜃) < 0, Δ2 = || |
|
| 0 − b2 (2 − 𝜃) || 8. Cruz JM (2013) Modeling the relationship of globalized supply
|
| −b (2 − 𝜃) | chains and corporate social responsibility. J Clean Prod 56:73–85
| 1 0 0 |
| | 9. Dutta P, Das D, Schultmann F, Frhling M (2016) Design and plan-
= b1 b2 (2 − 𝜃)2 >0 and Δ3 = || 0 − b1 (2 − 𝜃) 0 |
|
| | ning of a closed-loop supply chain with three way recovery and
| 0 0 − b (2 − 𝜃) | buy-back offer. J Clean Prod 135:604–619
| 3 |
= −b1 b2 b3 (2 − 𝜃)3 < 0. Clearly, Δn = (−1)n b1 b2 ...bn (2 − 𝜃)n . 10. Ernst CG, Young Overall Report Oct 2001 – Oct 2002. Online
availability of public services: how does Europe progress? Web
Δn > 0 if n is even and Δn < 0 if n is odd . Now, Based Survey On Electronic Public Services. https​://www.capge​
| −b1 (2 − 𝜃) 0 0 − − b1 T || mini.be/pdf/CGEYE​urope​Onlin​ePubl​icSer​vices​Overa​llRep​ort
|
| | 11. Ghobakhloo M, Tang SH, Zulkifli N, Ariffin MKA (2013) An inte-
| 0 − b2 (2 − 𝜃) 0 ... ... b2 T |
| | grated framework of green supply chain management imple-
| b3 T ||
| 0 0 − b3 (2 − 𝜃) ... ...
mentation. Int J Innov Manag Technol 4(1):86
Δn+1 = |Hc | = | |
| − − − − − − || 12. Govindan K, Soleimani H, Kannan D (2015) Reverse logistics and
|
| |
| − − − − − − | closed-loop supply chain: a comprehensive review to explore
| |
| − 2𝛽 || the future. Eur J Oper Res 240(3):603–626
| b1 T b2 T b3 T ... ...
13. Guide VDR, Wassenhove LNV (2003) Business aspects of closed-
O n s i m p l i f i c a t i o n w e h a v e , Δn+1 = (−1)n loop supply chains. Carnegie Mellon University Press, Pittsburgh
∑n
[2b1 b2 … bn 𝛽(2 − 𝜃)n + b1 b2 … bn T 2 (2 − 𝜃)(n−1) ( i=1 bi )] . 14. Hosseini-Motlagh SM, Nouri-Harzvili M, Choi TM, Ebrahimi S
Hence, E(vc ) will be concave if Δn+1 > 0 when n is odd and (2019) Reverse supply chain systems optimization with dual
channel and demand disruptions: sustainability, CSR investment
Δn+1 < 0 when n is even. On other words, E(vc ) will be con- and pricing coordination. Inf Sci 503:606–634
cave if the Hessian matrix, E(vc ) is negative definite, i.e., 15. Kazemi N, Modak NM, Govindan K (2018) A review of reverse
the eigenvalues of the matrix are all negative real num- logistics and closed loop supply chain management studies
bers. From the above discussion we have the following published in IJPR: a bibliometric and content analysis. Int J Prod
Res 57(15–16):4937–4960
proposition. 16. Kopel M, Brand B (2012) Socially responsible firms and endog-
enous choice of strategic incentives. Econ Model 29:982–989
17. Ma C (2015) Research progress of closed-loop supply chain
Appendix 2 based on recycling remanufacturing. Knowl Econ 4:104–106
18. Mafakheri F, Nasiri F (2013) Revenue sharing coordination in
reverse logistics. J Clean Prod 59:185–196
Differentiating optimal whlesale price and selling prices of 19. Modak NM, Kazemi N, Cardenas-Barron LE (2019) Investigating
the decentralized channel with respect to 𝜃 we get, structure of a two-echelon closed-loop supply chain using social

Vol.:(0123456789)
Research Article SN Applied Sciences (2019) 1:1189 | https://doi.org/10.1007/s42452-019-1173-1

work donation as a Corporate Social Responsibility practice. Int 33. Ramos TRP, Gomes MI, Barbosa-Povoa AP (2014) Planning a sus-
J Prod Econ 207:19–33 tainable reverse logistics system: balancing costs with environ-
20. Modak NM, Modak N, Panda S, Sana SS (2018) Analyzing struc- mental and social concerns. Omega 48:60–74
ture of two-echelon closed-loop supply chain for pricing, quality 34. Saha S, Sarmah SP, Moon I (2016) Dual channel closed-loop sup-
and recycling management. J Clean Prod 171:512–528 ply chain coordination with a reward-driven remanufacturing
21. Modak NM, Panda S, Sana SS (2016a) Two-echelon supply chain policy. Int J Prod Res 54(5):1503–1517
coordination among manufacturer and duopolies retailers with 35. Salvi BL, Jindal S (2019) Recent developments and challenges
recycling facility. Int J Adv Manuf Technol 87(5–8):1531–1546 ahead in carbon capture and sequestration technologies. SN
22. Modak NM, Panda S, Sana SS (2016b) Pricing policy and coor- Appl Sci 1(8):885
dination for a two-layer supply chain of duopolistic retailers 36. Savaskan RC, Bhattacharya S, Van Wassenhove LN (2004) Closed-
and socially responsible manufacturer. Int J Logist Res Appl loop supply chain models with product remanufacturing.
19(6):487–508 Manag Sci 50(2):239–252
23. Modak NM, Panda S, Sana SS, Basu M (2014) Corporate social 37. Schrady DA (1967) A deterministic inventory model for repara-
responsibility, coordination and profit distribution in a dual- ble items. Nav Res Logist (NRL) 14(3):391–398
channel supply chain. Pac Sci Rev 16(4):235–249 38. Soni U, Roy A, Verma A, Jain V (2019) Forecasting municipal
24. Nematollahi M, Hosseini-Motlagh SM, Ignatius J, Goh M, Nia MS solid waste generation using artificial intelligence modelsa case
(2018) Coordinating a socially responsible pharmaceutical sup- study in India. SN Appl Sci 1(2):162
ply chain under periodic review replenishment policies. J Clean 39. Srivastava SK, Srivastava RK (2006) Managing product returns for
Prod 172:2876–2891 reverse logistics. Int J Phys Distrib Logist Manag 36(7):524–546
25. Ni D, Kevin LW, Tang X (2010) Social responsibility allocation in 40. Vickers J (1985) Delegation and the theory of the firm. Econ J
two-echelon supply chains: insights from wholesale price con- 95:138–147
tracts. Eur J Oper Res 207(3):1269–1279 41. Wang X, Wang YH (2013) Coordination and pricing for closed-
26. Pan YM, Shan XL, Duan YX (2015) Cost-benefit optimization loop supply chain based on demand and collection disruptions.
analysis of reverse logistics. Value Eng 14:12–13 Comput Integr Manuf Syst 19(3):624–630
27. Panda S, Modak NM (2016) Exploring the effects of social 42. Wei J, Wei H (2011) A study on the reclaim cooperation with the
responsibility on coordination and profit division in a supply third-party reverse logistics enterprises. J Syst Manag 6:702–709
chain. J Clean Prod 139:25–40 43. Yi YY (2010) Study on closed-loop supply chain models under
28. Panda S, Modak NM, Basu M, Goyal SK (2015) Channel coordina- different market power. J Syst Manag 19(4):389–396
tion and profit distribution in a social responsible three-layer 44. Yoo SH, Kim D, Park MS (2015) Pricing and return policy under
supply chain. Int J Prod Econ 168:224–233 various supply contracts in a closed-loop supply chain. Int J Prod
29. Panda S, Modak NM, Pradhan D (2016) Corporate social respon- Res 53(1):106–126
sibility, channel coordination and profit division in a two-eche- 45. Zheng Y, Shu T, Wang S, Chen S, Lai KK, Gan L (2018) Analysis of
lon supply chain. Int J Manag Sci Eng Manag 11(1):22–33 product return rate and price competition in two supply chains.
30. Panda S, Modak NM, Cardenas-Barron LE (2017a) Coordination Oper Res 18(2):469–496
and benefit sharing in a three-echelon distribution channel with 46. Zheng J, Wu J (2016) Research on closed-loop supply chain with
deteriorating product. Comput Ind Eng 113:630–645 the third party recycler based on consumer preference. J Serv
31. Panda S, Modak NM, Cardenas-Barron LE (2017b) Coordinating Sci Manag 9(02):175
a socially responsible closed-loop supply chain with product
recycling. Int J Prod Econ 188:11–21 Publisher’s Note Springer Nature remains neutral with regard to
32. Pedersen ER, Andersen M (2006) Safeguarding corporate social jurisdictional claims in published maps and institutional affiliations.
responsibility (CSR) in global supply chains: how codes of con-
duct are managed in buyer–supplier relationships. J Public Aff
6(3–4):228–240

Vol:.(1234567890)