Sustainability 1406486

Article 1
Optimal Pricing and Green Product Design Strategies in a Sus- 2
tainable Supply Chain Considering Government Subsidy and 3
Different Channel Power Structures 4
Kelei Xue 1, Guohua Sun 1,*, Yuyan Wang 1,* and Shuiye Niu 2 5
1 School of Management Science and Engineering, Shandong University of Finance and Economics, Jinan 6
250014, China; keleixue@gmail.com (K.X.); nksungh@qq.com (G.S.); wangyuyan1224@126.com (Y.W.) 7
2 School of Innovation and Entrepreneurship, Shandong University, Qingdao 266237, China; blueangle- 8
love@126.com (S.N.) 9
* Correspondence: nksungh@qq.com (G.S.); wangyuyan1224@126.com (Y.W.) 10
Abstract: Green product design is a crucially important measure to support sustainable develop- 11
ment in circular economy era. In this paper, we study the multi-product pricing and green product 12
design strategies under different supply chain structures and government subsidy strategies. Con- 13
sidering different channel leadership, we establish the centralized (C), manufacturer-led (MS) and 14
retailer-led (RS) supply chain models, respectively. By applying a game-theoretical approach, cor- 15
responding equilibrium pricing, green product design and government’s subsidy decisions under 16
different supply chain structures are obtained. Through comparison and numerical analysis, we 17
find that: (1) the different subsidy strategies of the government have important impact on the de- 18
velopment of the green product. When the government provides a uniform subsidy strategy, a RS 19
supply chain can bring greener product, more market demands, more profit and more social wel- 20
Citation: Xue, K.; Sun, G.; Wang, Y.;
fare. (2) When the government provides a differentiated subsidy strategy, MS and RS supply chain 21
Niu, S. Optimal Pricing and Green
structures can both bring greener product and more market demands for green product than the 22
Product Design Strategies in a Sus-
centralized supply chain. They can also bring the same social welfare and the same product in green 23
tainable Supply Chain Considering
Government Subsidy and Different
design level. However, MS supply chain structure can bring more profit for the firm. (3) The con- 24
Channel Power Structures. Sustaina- sumers’ green awareness has a positive impact on the design and development of the green product. 25
bility 2021, 13, x. It is beneficial for the firm to adopt reasonable measures to boost environmental awareness of con- 26
https://doi.org/10.3390/xxxxx sumers in order to realize the sustainable development of our society. 27
Academic Editor(s): Keywords: green product design; multi-product pricing; government subsidy; channel power 28
structures 29
Received: date 30
Accepted: date
Published: date
1. Introduction 31
Publisher’s Note: MDPI stays neu-
tral with regard to jurisdictional
Sustainable supply chain management has received increasing attention in circular 32
claims in published maps and institu- economy era, driven by environment responsibilities, consumer pressures, government 33
tional affiliations. regulations and competitor pressures [1]. Many scholars study green product design, re- 34
manufacturing and reverse logistics, etc., to promote sustainable supply chain manage- 35
ment [2]. Thereinto, green product design plays a crucial role in maximizing firm’s profits 36
and minimizing environmental impact. Green design provides many obvious benefits for 37
Copyright: © 2021 by the authors. enterprises, including: increasing the possibility of additional sales, increasing consum- 38
Submitted for possible open access ers’ satisfaction, increasing social responsibility and environmental performance, abiding 39
publication under the terms and con-
by current and future regulations, improving brand reputation. Nowadays, a growing 40
ditions of the Creative Commons At-
number of consumers are inclined to buy environmentally friendly and pollution-free 41
tribution (CC BY) license (http://crea-
green products [3]. In a global survey conducted by Accenture, more than 80% of 42
tivecommons.org/licenses/by/4.0/).
Sustainability 2021, 13, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/sustainability

Sustainability 2021, 13, x FOR PEER REVIEW 2 of 27
interviewees show their high green preference when making purchasing choices 1. Carbon 43
Trust surveys also show that nearly one-fifth of customers prefer green products even 44
though these products are more expensive than regular ones2. Consumers’ green purchas- 45
ing behaviors are directly related to consumers’ environmental awareness (CEA), which 46
is a key market-driven factor for consumers’ purchasing options [4]. 47
Our research is mainly motivated by the following enterprises’ cases. Many enter- 48
prises are expanding their market share by launching green versions of traditional brown 49
products, which are very similar in all aspects except for the green characteristic. In the 50
automotive industry, Toyota provides hybrid electric versions of vehicles (e.g., Camry, 51
Carola, Prius) which are eco-friendly, in addition to traditional models that use gasoil. 52
Volkswagen promotes and develops plug-in hybrid electric vehicles (e.g., Passat, Audi 53
A3, A6, Q5, Q7, etc.) which have large battery capacity and are more eco-friendly than the 54
traditional models. Audi brand also provides pure electric version of vehicles in Q2 model 55
(i.e., Q2L e-tron) which can be regarded as the most environmentally friendly car com- 56
pared with the above two models. Similar situations can be observed in other industries. 57
Almost all home appliances brands (e.g., Siemens, Haier, Gree) produce green versions of 58
products, which are similar to traditional products in all aspects, but differ in terms of 59
energy consumption. In the apparel industry, many large brands (e.g., H&M, Zara, Nike), 60
provide green versions of clothings, which are similar to ordinary clothings in all aspects, 61
but are made of 100% organic fibers. Moreover, enterprises often have different channel 62
powers and are in different positions in the supply chain. In the previous examples, 63
Toyota, Siemens and Haier have more channel powers compared with their retailers. On 64
the contrary, H&M, Zara and Nike have the channel leadership compared with their sup- 65
pliers. Another similar example is Suning. Suning, a giant retailer, supports and cooper- 66
ates with its upstream manufacturer (e.g., ANGEL) to improve product greening level 3. 67
In reality, many governments, such as China, America, India, etc., have provided 68
various subsidy policies to promote green innovation and sustainable development. For 69
example, in 2012, China decided to arrange financial subsidy at 13% of the sales price 70
(¥26.5 billion in total) to start the promotion of home appliances (e.g., air conditioner, re- 71
frigerator, washing machine, etc.) that meet energy-saving standards. In America, con- 72
sumers who purchased new energy vehicles will enjoy tax credits ranging from $2500 to 73
$7500. Meanwhile, China has also launched its own series of new energy vehicle subsidy 74
policies. In 2019, China’s subsidy standard for pure electric vehicles with a driving range 75
of more than 400 kilometers was ¥25000 per vehicle. The government of India Ministry of 76
Power subsidized the purchasing of LED lightbulbs in the country through the UJALA 77
program4. The underlying rationale is still unclear on the government subsidy associated 78
with the expected behavior from enterprise. 79
Based on the aforementioned firms’ cases in practice, we propose the following re- 80
search questions: 81
Research Question 1: With the coexistence of the regular and green products, how 82
should the firm formulate reasonable pricing and green product design strategies when 83
taking CEA into consideration? 84
Research Question 2: What are the effects of the different government subsidy strat- 85
egies, i.e., uniform subsidy and differentiated subsidy, on the firm’s multi-product pricing 86
and green product design? 87
Research Question 3: At the supply chain level, if it exists, what is the optimal chan- 88
nel structure for the firm from the perspectives of firm’s profit and total social welfare? 89
Our research problem is related to the literature about multi-product pricing, green 90
product design and government policy under sustainable supply chains. Our paper is 91
1 Available at: https://newsroom.accenture.com/article_display.cfm?article_id=4801, accessed on 11 July 2021.

2 Available at: http://www.carbontrust.com/news/2011/07/consumer-dem and-for-lower-carbon-lifestyles-is-putting-
pressure-on-business, accessed on 11 July 2021.
3 Available at: http://www.cinn.cn/gysj/201804/t20180416_180606_wap.html, accessed on 11 July 2021.
4 Available at: http://www.ujala.gov.in/, accessed on 11 July 2021.

closely related to Agi and Yan [5], Hong et al. [6] and Bian et al. [7]. Agi and Yan [5] in- 92
vestigates pricing and positioning strategies for brown and green products under differ- 93
ent channel power structures. However, their work doesn’t involve the issues of green 94
product design and government subsidy. Hong et al. [6] investigate the effects of tax reg- 95
ulation, consumers' reference behaviors and CEA on the green product design and com- 96
petitive pricing strategies under both manufacturer-led and retailer-led supply chain 97
structures. The main difference is that our study considers two types of consumer subsidy 98
strategies (not tax regulation) from government and investigates their effects on the green 99
product design. We also find that manufacturer-led and retailer-led supply chains have 100
the same social welfare when the government provides a differentiated subsidy strategy, 101
which is different from their research. Bian et al. [7] is one of the few studies to explore 102
the impact of the environmental subsidy policies (i.e., manufacturer vs. consumer sub- 103
sidy) from government on the green product design. Different from their study, our paper 104
considers the other two types of consumer subsidies: uniform subsidy and differentiated 105
subsidy. However, our paper pays more attention to the game relationship between the 106
supply chain members and tries to explore how it affects the green product design and 107
multi-product pricing strategies. 108
To answer the above research questions and fill these research gaps, we investigate 109
the multi-product pricing and green product design strategies in a sustainable supply 110
chain considering government’s different subsidy strategies and different channel power 111
structures. Firstly, we construct the consumer demand function by the utility model. Af- 112
terwards, considering different channel powers and government’s two types of subsidy 113
strategies, we establish the Stackelberg game models in centralized, manufacturer-led and 114
retailer-led supply chain, respectively. By applying a game-theoretical approach, corre- 115
sponding equilibrium pricing, green product design and government’s subsidy decisions 116
under different supply chain structures are obtained. Finally, we compare the optimal 117
decisions, profits of the supply chain and total social welfare under different channel 118
power structures and subsidy strategies of government. 119
To sum up, our main contributions in this paper are as follows. 120
1. To the best of our knowledge, our paper is one of first to investigate the joint deci- 121
sions of pricing and green product design strategies considering the coexistence of 122
the regular product and green product under different channel power structures and 123
subsidy strategies of government. 124
2. Our paper considers the government’s two types of consumer subsidy strategies: uni- 125
form subsidy and differentiated subsidy, which are seldomly studied by relevant 126
scholars. Under the differentiated subsidy strategy, the government can decide its 127
optimal subsidy based on different supply chain structures, so as to maximize the 128
total social welfare consisting of monetary welfare and environmental welfare. 129
3. Under the changing operational environment, the optimal decisions and profits of 130
the supply chain and the optimal government subsidies and total social welfare un- 131
der different supply chain structures are compared and analyzed so that more con- 132
clusions and managerial implications can be derived. 133
The rest of the study is structured as follows. We give the literature review in Section 134
2. In Section 3, we describe the problem and the model. In Section 4, we formulate the 135
optimization problems of the supply chain and the government, and give the correspond- 136
ing equilibrium decisions under three different power structures. In Section 5, we com- 137
pare the optimal decisions of the supply chain and the government under two types of 138
government subsidy strategies. Furthermore, through numerical analysis method, more 139
managerial implications are derived. Finally, Section 6 concludes the study and presents 140
future research directions. 141
142
2. Literature Review 143

Our paper is closely related to three streams of literatures: multi-product pricing, 144
green product design and government policy in sustainable supply chains. In this section, 145
we review literatures relevant to each stream and highlight the differences between the 146
existing research and our work. 147
2.1. Multi-product Pricing 148

Multi-product pricing is one of the most significant topics drawing attention from 149
relevant scholars from green supply chain management [8-18]. Our paper considers pric- 150
ing scenario with the coexistence of the regular and green products. In the following, we 151
mainly review papers related to this scenario. 152
Su et al. [8] investigate the green product’s pricing and environmental quality design 153
problems and derive conditions under which Market Segmentation strategy (two prod- 154
ucts are offered) or Mass Marketing strategy (a single product is offered) is adopted by 155
the firm. By establishing and solving a two-product optimization model, Li and Li [9] also 156
study green market segmentation and their pricing decisions, where the government sub- 157
sidy regulation is considered in their work. Under consumers’ market segmentation and 158
different channel power structures, Agi and Yan [5] investigates pricing and positioning 159
strategies for brown and green products and examined the coordination mechanisms in 160
two decentralized supply chains. 161
Some scholars investigate the pricing problem in competitive supply chain. Li and Li 162
[10] consider the competition at the supply chain level and examine how the sustainability 163
competition influences the product sustainability. Deng et al. [12] also consider the com- 164
petition between two supply chains and explore how the horizontal competition and ver- 165
tical competition affect product sustainability and supply chain profits. The above two 166
studies both assume that the consumers’ demand function is sustainability-sensitive. 167
From the perspective of economic and environment, Sim et al. [13] investigate the effects 168
of supply chain competition on the abatement efforts and eco-friendly social welfare. 169
Their results show that the most efficient market structures may be not efficient from a 170
broader perspective. Under an eco-label policy, Gao et al. [14] study the pricing and coor- 171
dination problem in a dual-channel green supply chain. 172
There are also scholars exploring the effects of CEA on the optimal pricing strategies. 173
By using a multi-product newsvendor model, Zhang et al. [15] investigate the green prod- 174
uct pricing problem and explore the effects of CEA on order quantities and channel coor- 175
dination. Considering CEA and non-green product reference, Hong et al. [16] investigate 176
the green-product pricing problem under three scenarios: single-product pricing, dual- 177
product competition, and asymmetric-information case. Their analysis results show that 178
when facing consumers with different buying behaviors, a differential pricing strategy 179
should be adopted. Under three different market scenarios, Zhang et al. [17] study the 180
introduction problem of green product and strategic pricing, and examine how the above 181
decisions are affected by CEA. 182
Unlike our paper, all the above papers don’t consider the investment decisions of the 183
supply chain members for green product design or development for environmental sus- 184
tainability under different supply chain channel structures. 185
2.2. Green Product Design 186

Green product design is an effective method for realizing a circular economy and has 187
been widely studied by numerous scholars from a variety of perspectives [18-28]. In the 188
following, we mainly review papers which study green product design issue from the 189
perspectives of cooperation, competition and government regulation. 190
Some scholars study the cooperation contracts or strategies which are used to en- 191
hance green product design. Considering two types of cost sharing contracts, Ghosh and 192
Shah [19] explore the coordination issues and demonstrate the impacts of cost sharing 193
contracts on product greening levels, prices, and profits within the supply chains. Hong 194
and Guo [20] study three types of coordination contracts, i.e., price-only, green-marketing 195
cost-sharing, and two-part tariff contracts, and find that cooperation among partners can 196
help the supply chain realize environmental improvements. Dong et al. [21] study who 197
should lead and invest in green product development (GPD) in a supply chain consisting 198
of one manufacturer and one retailer. By developing a two-period model, they find that 199
when the manufacturer invests in GPD, the products are greener and the manufacturer 200
earns more. In a decentralized supply chain, Li et al. [22] consider two types of cooperative 201
contracts: contracting designing (CD) format and contracting marketing format (CM), and 202
explore how to select appropriate cooperative contract to boost product greening level. 203
Some other scholars investigate the effects of competition on the green product de- 204
sign strategies. Zhu and He [23] explore the firms’ green product design issues under sup- 205
ply chains competition and address how the green product types, the types of competition 206
(price competition or greenness competition) affect “greenness” of products. In the case 207
of considering consumer returns, Guo et al. [24] investigate how the greenness competi- 208
tion affects green product design in fashion apparel. Considering different channel pow- 209
ers, Ma et al. [25] explore how the competition between upstream manufacturers and sup- 210
ply chain structure affects the green manufacturing level. Considering the competition 211
between the upstream firms of the supply chain, Du et al. [26] also explore the competition 212
between the upstream manufacturers and how this type of competition affects the green 213
product design strategy adoption and pricing strategy. Unlike our paper, the above liter- 214
atures don’t involve the issue of the government subsidy and different channel structures. 215
2.3. Government Policy 216

There are also scholars exploring the effects of the government regulation or subsidy 217
on the green product design strategies [29-38]. Cai et al. [32] explore three types of envi- 218
ronmental taxes: linear tax, constant tax and zero tax, and examine how the producer’s 219
optimal DfE (Design for Environment) level could be influenced by them. Yalabik and 220
Fairchild [35] study how the regulatory penalty and retail competition influence the in- 221
vestment in emission reduction. Considering the extended producer responsibility (EPR) 222
policy, Subramanian et al. [36] investigate how the EPR regulation affects a durable prod- 223
uct design. Their study takes into account the tax policy in the green product design. Con- 224
sidering consumers’ reference behaviors, Hong et al. [6] investigate the effects of tax reg- 225
ulation, consumers' reference behaviors and CEA on the green product design and pricing 226
strategies. Different from their study, we consider the government’s two different subsidy 227
strategies and investigate their effects on the green product design. In the textiles and 228
apparel supply chains, Shen et al. [37] explore the effects of environmental taxes on the 229
green technology adoption. In a research joint venture (RJV), Chen et al. [38] endogenize 230
government subsidy and explore effects of two types of subsidy and two forms of RJV 231
formation on the level of sustainability innovation. Bian et al. [7] consider two types of 232
government technology subsidy policies: consumer and manufacturer subsidies, and ex- 233
plore the effects of different subsidy policies on the manufacturer’s green investment de- 234
cisions. Different from our study, the above-mentioned papers don’t consider the effects 235
of a differentiated subsidy strategy of government on the green product design under 236
different channel power structures. 237
2.4. Summary 238

In order to show the novelty of this article more clearly, we give the Table 1 to present 239
the main differences between our work and related studies, which can be summarized in 240
three points. Firstly, most papers on pricing don’t consider green product design and gov- 241
ernment subsidy, or they don’t explore the impacts of different channel powers structures 242
on the green product design. Secondly, most papers on green product design don’t in- 243
volve the issues of the government subsidy and different channel structures. Thirdly, 244
there are few scholars who have studied the government’s two types of consumer subsidy 245
strategies: uniform subsidy and differentiated subsidy. To fill the above research gaps, we 246
explore the multi-product pricing and green product design strategies in a sustainable 247
supply chain considering government’s different subsidy strategies and different channel 248
power structures. 249
Moreover, our paper adopts the game-theoretical approach. The reasons are as fol- 250
lows. Most relevant scholars (refer to Table 1) adopt a game-theoretical approach to model 251
related research problems. A few scholars use optimization or other modeling methods. 252
For example, Zhang et al. [15] use a newsvendor model method in a random demand 253
environment. Considering a single decision maker, Hong et al. [16] and Cai et al. [32] both 254
adopt an optimization modeling method. Our paper examines the strategic interactions 255
among multiple decision makers. Therefore, we adopt a game-theoretical approach to 256
model and analyze our problem. 257
Table 1. Main differences between our work and related studies. 258
Issues involved Decision variables

Channel Modeling Government
Articles Green
structure method policy CEA Product type Price
design
Subsidy Tax
Li and Li [9] One firm Game √ √ Regular+Green √
Li and Li [10] Two chains Game Green √
Agi and Yan [5] MS, RS, C Game √ Regular+Green √
Deng et al. [12] Two chains Game Green √
Sim et al. [13] Two chains Game Green √ √
Gao et al. [14] Dual-channel Game √ Green √
Zhang et al. [15] MS, C Newsvendor √ Regular+Green √
Optimization,
Hong et al. (2018) [8] Two firms √ Regular+Green √
Bayes
Zhang et al. [17] MS, C Game √ √ Regular+Green √
Ghosh and Shah [19] MS, C Game √ Green √ √
Hong and Guo [20] MS, C Game √ Green √ √
Dong et al. [21] MS Two-period √ Regular+Green √ √
Zhu and He [23] MS, C Game √ Green √ √
Multi-
Guo et al. [24] MS √ Green √ √
methodological
Du et al. [26] MS Game √ Green √ √
Hong et al. [6] MS, RS Game √ √ Regular+Green √ √
Shen et al. [37] MS Game √ Green √ √
Cai et al. [32] One firm Optimization √ Green √ √
Bian et al. [7] M+G Game √ √ Green √ √
Our paper MS, RS, C Game √ √ Regular+Green √ √
1MS: Manufacturer-led Supply chain; RS: Retailer-led Supply chain; C: Centralized Supply chain; G+M: Manufacturer and 259
Government; Regular+Green: Regular (Non-green) and Green products; CEA: Consumers’ Environmental Awareness. 260
3. Model Descriptions 261

Consider a decentralized supply chain that consists of a manufacturer and a retailer. 262
The manufacturer produces two types of products: regular product (labeled as product 1) 263
and green product (labeled as product 2). Green product can be regarded as green ver- 264
sions of regular product, with same or similar product function. For example, Toyota and 265
Volkswagen simultaneously provide traditional fuel vehicles and hybrid electric vehicles 266
to the market in the same models (e.g., Camry, Carola, Passat, Audi A3, etc.). Compared 267
with the regular product, green product is designed to has less environmental impact at 268
different stages of its life cycle, including manufacturing, usage and recycle stage. We use 269
𝑒 (0 < 𝑒 < 1) to denote the green design level of the green product. The retailer whole- 270
sales two types of products from the manufacturer and sells them to consumers in the 271
market. 272
The manufacturer needs to invest in research and development (R&D) costs to de- 273
1
velop environmentally friendly products. The amount of capital investment is 𝑘𝑒 2 , 274
2
where 𝑘 > 0 denotes the R&D cost coefficient for the green product. This quadratic cost 275
function is widely used in relevant literature, e.g., [21,23,39,40]. The production cost per 276
unit of green product is 𝑐2 = 𝑐1 + 𝑐, where 𝑐1 is the production cost per unit of the regu- 277
lar product. Then 𝑐 > 0 can be looked as the increase in production cost of the green 278
product compared with the regular product. In general, the production of the green prod- 279
uct is usually more expensive than that of the regular product [41,42]. For example, the 280
production cost of coffee with environmentally friendly production technology has in- 281
creased by 30%5. Without loss of generality, we assume 𝑐1 = 0. When 𝑐1 = 0, parameter 282
𝑐 can be looked as the production cost of the green product. 283
Consumers are environmentally conscious and homogeneous in obtaining environ- 284
mental utility from the green product, but are heterogeneous in terms of their evaluations 285
of the functional attributes of products [6,16,20]. We use 𝑉 to denote the utility of con- 286
sumers obtaining from the functional attribute of the product. We assume 𝑉 is uniformly 287
distributed in the interval [0, 1] and its probability density function (PDF) and cumulative 288
distribution function (CDF) are 𝑓(∙) and 𝐹(∙), respectively [43,44]. The function formulas 289
of 𝑓(∙) and 𝐹(∙) are given by Eq. (1). The utility of consumers purchasing the regular 290
product is 𝑈1 = 𝑉 − 𝑝1 , where 𝑝1 is the retail price of the regular product. 291
𝑓(𝑥) = 1
{ , 𝑥 𝑖𝑠 𝑡ℎ𝑒 𝑟𝑒𝑎𝑙𝑖𝑧𝑒𝑑 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑉. (1)
𝐹(𝑥) = 𝑥
The utility of a consumer deriving from the green product’s functional attribute is 292
𝜃𝑉, where 𝜃 > 1 indicate that the consumers have higher willingness to pay for the func- 293
tional attribute of the green product [16]. In reality, generally speaking, the green version 294
of regular product usually has higher functional attribute compared to the regular prod- 295
uct, e.g., air conditioner, etc. Then, we assume that 𝜃 > 1. We also solve our models when 296
𝜃 < 1. We find that our main findings and managerial insights are robust. The consumer 297
can also homogeneously derive utility from the environmental attribute of the green prod- 298
uct. The environmental utility is assumed to be 𝜆𝑒 [45,46], where 𝜆 > 0 is the consumers’ 299
preference coefficient for the green product’s greenness and can also indicate the level of 300
CEA [47]. In practice, the government can provide subsidy to the consumers who buy the 301
green product. The utility of consumers purchasing the green product is 𝑈2 = 𝜃𝑉 − 𝑝2 + 302
𝑠 + 𝜆𝑒, where 𝑠 denotes the government subsidy to the consumer who purchases a green 303
product and 𝑝2 is the retail price of the green product. We don’t consider the heteroge- 304
neity of consumers in purchasing quantity, then we assume one consumer purchases at 305
most one unit of the product. The market size is normalized to 1. 306
In the market, the consumers decide which type of product to buy by comparing their 307
utilities. The consumers buy the regular product under the condition of 𝑈1 ≥ 𝑈2 , 𝑈1 ≥ 0, 308
𝑝 −𝑝 −𝑠−𝜆𝑒
i.e., 𝑝1 ≤ 𝑉 ≤ 2 1 . The consumers buy the green product if and only if 𝑈2 ≥ 𝑈1 , 309
𝜃−1
𝑝2 −𝑠−𝜆𝑒 𝑝2 −𝑝1 −𝑠−𝜆𝑒
𝑈2 ≥ 0, i.e., max { , } ≤ 𝑉 ≤ 1. Similar to Hong et al. [6] and Hong et al. 310
𝜃 𝜃−1
[16], we only consider the situation where the regular and green products are both pre- 311
sented in the market. Let 𝑞1 and 𝑞2 denote the number of consumers buying the regular 312
and green products, respectively. Then, the market demands for the regular and green 313
products are 314
5 Available at: http://www.triplepundit.com/2014/11/economics-sustainable-coffee-production/, accessed on 11 July,

2021.
𝑝2 −𝑝1 −𝑠−𝜆𝑒
𝜃−1 𝑝2 − 𝜃𝑝1 − 𝑠 − 𝜆𝑒
𝑞1 = ∫ 𝑑𝐹(𝑥) = , (2)
𝑝1 𝜃−1
1
𝑝2 − 𝑝1 − 𝑠 − 𝜆𝑒
𝑞2 = ∫ 𝑑𝐹(𝑥) = 1 − . (3)
𝑝2 −𝑝1 −𝑠−𝜆𝑒 𝜃−1
𝜃−1
Accordingly, the profits of the manufacturer, retailer, and supply chain are 315
1
𝜋𝑚 (𝑤1 , 𝑤2 , 𝑒) = 𝑤1 𝑞1 + (𝑤2 − 𝑐)𝑞2 − 𝑘𝑒 2 , (4)
2
𝜋𝑟 (𝑝1 , 𝑝2 ) = (𝑝1 − 𝑤1 )𝑞1 + (𝑝2 − 𝑤2 )𝑞2 , (5)
1
𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) = 𝑝1 𝑞1 + (𝑝2 − 𝑐)𝑞2 − 𝑘𝑒 2 . (6)
2
In Eq. (4), the first two terms are the manufacturer’s profits from selling the regular 316
and green products to the retailer; the third term is R&D investment for the green product. 317
In Eq. (5), the first and second terms denote the retailer’s profits from selling the regular 318
and green products to the consumers. Eq. (6) is the sum of the Eq. (4) and Eq. (5). 319
Table 2. The notations used in this paper. 320
Notations Descriptions
production cost of product i, i={1, 2}, “1” and “2” denote the
𝑐𝑖 , 𝑐
regular and green products, respectively, 𝑐2 = 𝑐1 + 𝑐.
𝑘 the green product R&D cost coefficient
𝜃 green product function coefficient
𝜆 consumers’ green preference coefficient
𝑉, 𝑓(∙), 𝐹(∙) the consumers’ willingness-to-pay (WTP) and its PDF, CDF
𝑡 negative environmental impact coefficient
𝑞𝑖 the demand for product i, i={1, 2}
𝜋𝑗𝑘 the profit of 𝑗 in 𝑘 model, j={m, r, sc}, k={C, MS, RS}
𝑆𝑊 𝑘 social welfare in 𝑘 model, k={C, MS, RS}
Decision variables
𝑒𝑘 green product design level in 𝑘 model, k={C, MS, RS}
𝑘
𝑠 government subsidy in 𝑘 model, k={C, MS, RS}
𝑤𝑖𝑘 wholesale price of product i in 𝑘 model, i={1, 2}, k={C, MS, RS}
𝑝𝑖𝑘 retail price of product i in 𝑘 model, i={1, 2}, k={C, MS, RS}
In addition, we also consider the impacts of supply chain decisions on entire social 321
welfare. Similar to Li and Li [9], the entire social welfare can be given by 322
𝑆𝑊 = 𝜋𝑠𝑐 + 𝐶𝑆 − 𝑠𝑞2 − 𝑡(𝑞1 + (1 − 𝑒)𝑞2 ). (7)

In Eq. (7), the first term 𝜋𝑠𝑐 is the profit of the total supply chain; the second term 323
𝐶𝑆 is consumer’s surplus; the third term 𝑠𝑞2 is the total subsidy given by government; 324
the fourth term 𝑡(𝑞1 + (1 − 𝑒)𝑞2 ) is the negative impacts (e.g., carbon emission) of the 325
regular and green products on the environment in entire life cycle, where 𝑡 denotes the 326
negative environmental impact coefficient. The first three terms of Eq. (7) are monetary 327
welfare, and the last term is environmental welfare. The consumer’s surplus 𝐶𝑆 can be 328
given by 329
1
𝜃−1
𝐶𝑆 = ∫ (𝑥 − 𝑝1 )𝑑𝐹(𝑥) + ∫ (𝜃𝑥 − 𝑝2 + 𝑠 + 𝜆𝑒)𝑑𝐹(𝑥) (8)
𝑝1
𝜃−1
To ensure that the mathematical model is solvable and avoid the trivial outcomes, 330
we make the following assumptions: 2𝑘(𝑐 − 𝑠) > 𝜆2 , 𝜃 − 1 > 𝑐 and 𝑘(𝜃 − 1) > 2𝜆(2𝑡 + 331
𝜆). For notational convenience, we use superscript “*”, “C”, “MS” and “RS” to denote 332
“optimal”, “centralized supply chain”, “manufacturer-led supply chain” and “retailer-led 333
supply chain”, respectively. Subscript “m”, “r” and “sc” refer to “manufacturer”, “re- 334
tailer” and “supply chain”, respectively. The model parameters and other notations used 335
in this paper are concluded in Table 2. 336
4. Model Analysis 337

In this section, considering different channel powers, centralized supply chain (C), 338
manufacturer-led supply chain (MS) and retailer-led supply chain (RS) models are estab- 339
lished. The game-theoretical method is used to analyze the three models. We also consider 340
the scenario where government can provide two different types of subsidy strategies, i.e., 341
uniform subsidy of government and differentiated subsidy of government. In each sub- 342
sidy strategy of government, we derive the equilibrium decisions of the supply chain and 343
government. 344
4.1. Centralized Supply Chain (C) Model 345

We first consider the scenario where the manufacturer and the retailer act as one 346
company. They make decisions from the perspective of whole supply chain. We call this 347
scenario as centralized supply chain (C) model. In the C model, the supply chain decides 348
the green design level of the green product and the prices of regular and green products. 349
Therefore, the supply chain’s optimization problem can be formulated as 350
𝐶 (𝑝
1 2
max 𝜋𝑠𝑐 1 , 𝑝2 , 𝑒) = 𝑝1 𝑞1 + (𝑝2 − 𝑐)𝑞2 − 𝑘𝑒 (9)
𝑝1 ,𝑝2 ,𝑒 2
where 𝑞1 and 𝑞2 are given by Eqs. (2) and (3). Solving the supply chain’s optimization 351
problem, we can derive the following Lemma 1. 352
353
Lemma 1. In the C model, the optimal retail prices and demands of the regular and green products, 354
1 2𝑘(𝜃−1)(𝑐+𝑠+𝜃)−(1+2𝑐)𝜆2
the optimal green product design level are 𝑝1𝐶∗ = , 𝑝2𝐶∗ = , 𝑞1𝐶∗ = 355
2 4𝑘(𝜃−1)−2𝜆2
2𝑘(𝑐−𝑠)−𝜆2 𝑘(𝑠+𝜃−𝑐−1) (𝑠+𝜃−𝑐−1)𝜆
, 𝑞2𝐶∗ = , and 𝑒 𝐶∗ = . 356
4𝑘(𝜃−1)−2𝜆2 2𝑘(𝜃−1)−𝜆2 2𝑘(𝜃−1)−𝜆2
The proofs of all Lemmas, Corollaries, Theorems and Propositions are in Appendix 357
A. According to Lemma 1, the following corollary can be obtained. 358
359
Corollary 1. In the C model, 360
𝜕𝑞1𝐶∗ 𝜕𝑞1𝐶∗ 𝜕𝑞1𝐶∗ 𝜕𝑞1𝐶∗ 𝜕𝑞1𝐶∗
(1) for regular product: > 0, > 0, < 0, < 0, < 0; 361
𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠
𝜕𝑝2𝐶∗ 𝜕𝑝2𝐶∗ 𝜕𝑝2𝐶∗ 𝜕𝑝2𝐶∗ 𝜕𝑝2𝐶∗ 𝜕𝑞2𝐶∗ 𝜕𝑞2𝐶∗
(2) for green product: < 0, > 0, > 0, > 0, > 0; < 0, < 0, 362
𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐
𝜕𝑞2𝐶∗ 𝜕𝑞2𝐶∗ 𝜕𝑞2𝐶∗ 𝜕𝑒 𝐶∗ 𝜕𝑒 𝐶∗ 𝜕𝑒 𝐶∗ 𝜕𝑒 𝐶∗ 𝜕𝑒 𝐶∗
> 0, > 0, > 0; < 0, < 0, > 0, > 0, > 0. 363
𝜕𝜃 𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠
According to Lemma 1 and Corollary 1, in the C model, no matter how the opera- 364
tional environment changes, the retail price of the regular product is a constant. The mar- 365
ket demand for the regular product increases in green product R&D cost coefficient, green 366
product production cost, and decreases in consumers’ green product function coefficient, 367
consumers’ green preference coefficient and government subsidy. In contrast, for the 368
green product, the retail price, market demand and green design level of the green prod- 369
uct decrease in green product R&D cost coefficient and green product production cost, 370
and increase in consumers’ green product function coefficient, consumers’ green prefer- 371
ence coefficient and government subsidy. 372
Now we turn to the government’s problem. The government’s goal is to maximize 373
the entire social welfare. The government can provide two subsidy strategies, i.e., uniform 374
subsidy of government and differentiated subsidy of government. Under the uniform sub- 375
sidy strategy, the government provide the same subsidy in the C, MS and RS models. 376
Under the differentiated subsidy strategy, the government provide different subsidies in 377
different models. In the following, we consider the scenario where the government pro- 378
vide differentiated subsidy. Under this scenario, the government first decides the optimal 379
subsidy policy for the green product. Then, it is followed by the supply chain’s optimiza- 380
tion problem. The government’s decision-making problem is formulated as 381
max 𝑆𝑊 𝐶 (𝑠)
= 𝐶∗
𝜋𝑠𝑐 + 𝐶𝑆 𝐶∗
− 𝑠𝑞2𝐶∗ − 𝑡(𝑞1𝐶∗ + (1 − 𝑒)𝑞2𝐶∗ ). (10)
𝑠
Substituting the optimal decisions given by Lemma 1 into Eq. (10), then we can obtain 382
the optimal subsidy by solving the government’s optimization problem. We use super- 383
script “D” to differentiate the optimal decisions under the differentiated subsidy strategy 384
of government. 385
Theorem 1. Under the differentiated subsidy strategy and C model, the government’s 386
optimal subsidy that maximizes the total social welfare is 387
(𝜃 − 1 − 𝑐)(𝑘(𝜃 − 1) + 2𝑡𝜆)
𝑠 𝐷𝐶∗ = (11)
𝑘(𝜃 − 1) − 𝜆(2𝑡 + 𝜆)
From Theorem 1, we can derive the following Corollary 2. 388
389
𝜕𝑠 𝐷𝐶∗ 𝜕𝑠 𝐷𝐶∗ 𝜕𝑠 𝐷𝐶∗ 𝜕𝑠 𝐷𝐶∗
Corollary 2. < 0, < 0, > 0, > 0. 390
𝜕𝑘 𝜕𝑐 𝜕𝜆 𝜕𝑡
391
When the government adopts a differentiated subsidy strategy, the increases of green 392
product R&D cost coefficient and green product production cost lead to the decrease of 393
the optimal government subsidy. While the increases of consumers’ green preference co- 394
efficient and negative environmental impact coefficient lead to the increase of the optimal 395
government subsidy. 396
Substituting the optimal government subsidy given by Eq. (11) into the optimal sup- 397
ply chain decisions given by Lemma 1, we can get Lemma 2 as follow. 398
Lemma 2. Under the differentiated subsidy strategy and C model, the optimal retail prices and 399
1
demands of the regular and green products, the optimal green product design level are 𝑝1𝐷𝐶∗ = , 400
2
𝑘(𝜃−1)(2𝜃−1)−(1+2𝑐)𝜆(2𝑡+𝜆) 1 𝑘(1+𝑐−𝜃) 𝑘(𝜃−1−𝑐)
𝑝2𝐷𝐶∗ = , 𝑞1𝐷𝐶∗ = + , 𝑞2𝐷𝐶∗ = , 𝑒 𝐷𝐶∗
= 401
2𝑘(𝜃−1)−2𝜆(2𝑡+𝜆) 2 𝑘(𝜃−1)−𝜆(2𝑡+𝜆) 𝑘(𝜃−1)−𝜆(2𝑡+𝜆)
(𝜃−1−𝑐)𝜆
. 402
𝑘(𝜃−1)−𝜆(2𝑡+𝜆)
4.2. Manufacturer-led Supply Chain (MS) Model 403

In this subsection, we consider the scenario where the manufacturer has more chan- 404
nel power than the retailer. We call this scenario as manufacturer-led supply chain (MS) 405
model, in which the manufacturer acts as the Stackelberg game leader, while the retailer 406
acts as the follower. In the MS model, the manufacturer first decides the green product 407
design level and the wholesale prices of the regular and green products. Then, the retailer 408
decides the retail prices of the regular and green products. The MS model is formulated 409
as 410
𝑀𝑆 (𝑤
1 2
max 𝜋𝑚 1 , 𝑤2 , 𝑒) = 𝑤1 𝑞1 + (𝑤2 − 𝑐)𝑞2 − 𝑘𝑒
𝑤1 ,𝑤2 ,𝑒 2
𝑝1∗ and 𝑝2∗ are derived from solving the following problem, 411
max 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) = (𝑝1 − 𝑤1 )𝑞1 + (𝑝2 − 𝑤2 )𝑞2

{ 𝑝1 ,𝑝2
where 𝑞1 and 𝑞2 are given by Eqs. (2) and (3). Solving the MS model, we can derive 412
Lemma 2 as follow. 413
Lemma 3. In the MS model, the optimal wholesale prices, retail prices and demands of the regular 414
1
and green products, and the optimal green product design level are 𝑤1𝑀𝑆∗ = , 𝑤2𝑀𝑆∗ = 415
2
4𝑘(𝜃−1)(𝑐+𝑠+𝜃)−(1+2𝑐)𝜆2 3 4𝑘(𝜃−1)(𝑐+3(𝑠+𝜃))−(3+4𝑐)𝜆2 4𝑘(𝑐−𝑠)−𝜆2

, 𝑝1𝑀𝑆∗ = , 𝑝2𝑀𝑆∗ = , 𝑞1𝑀𝑆∗ = , 416
8𝑘(𝜃−1)−2𝜆2 4 16𝑘(𝜃−1)−4𝜆2 4(4𝑘(𝜃−1)−𝜆2 )
𝑀𝑆∗ 𝑘(𝑠+𝜃−1−𝑐) (𝑠+𝜃−1−𝑐)𝜆
𝑞2 = , 𝑒 𝑀𝑆∗ = . 417
4𝑘(𝜃−1)−𝜆2 4𝑘(𝜃−1)−𝜆2
According to Lemma 3, the following Corollary 3 can be obtained. 418
Corollary 3. In the MS model, 419
𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗

(1) for the regular product: > 0, > 0, < 0, < 0, < 0; 420
𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠
𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑝2𝑀𝑆∗ 𝜕𝑝2𝑀𝑆∗

(2) for the green product: < 0, > 0, > 0, > 0; < 0, > 421
𝜕𝑘 𝜕𝑐 𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐
𝜕𝑝2𝑀𝑆∗ 𝜕𝑝2𝑀𝑆∗ 𝜕𝑞2𝑀𝑆∗ 𝜕𝑞2𝑀𝑆∗ 𝜕𝑞2𝑀𝑆∗ 𝜕𝑞2𝑀𝑆∗ 𝜕𝑞2𝑀𝑆∗ 𝜕𝑒 𝑀𝑆∗
0, > 0, >0 ; < 0, < 0, > 0, > 0, >0 ; < 422
𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠 𝜕𝑘
𝜕𝑒 𝑀𝑆∗ 𝜕𝑒 𝑀𝑆∗ 𝜕𝑒 𝑀𝑆∗ 𝜕𝑒 𝑀𝑆∗
0, < 0, > 0, > 0, > 0. 423
𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠
In the MS model, no matter how the operational environment changes, the wholesale 424
price and retail price of the regular product are both constants. Similar to the C model, the 425
market demand for the regular product increases in green product R&D cost coefficient, 426
green product production cost, and decreases in consumers’ green product function coef- 427
ficient, consumers’ green preference coefficient and government subsidy. For the green 428
product, the retail price, market demand and the green design level of the green product 429
decrease in green product R&D cost coefficient, and increase in consumers’ green product 430
function coefficient, consumers’ green preference coefficient and government subsidy. For 431
the wholesale price of the green product, it decreases in green product R&D cost coeffi- 432
cient, but increases in consumers’ green preference coefficient. The increase of green prod- 433
uct production cost leads to the increases of the wholesale price and retail price of the 434
green product, while it leads to the decreases of the market demand and green design 435
level of the green product. 436
Next, we turn to the government’s problem in the MS model. Similarly, under the 437
differentiated subsidy strategy, we consider the government’s problem of maximizing en- 438
tire social welfare. The government first decides the optimal subsidy for the green prod- 439
uct. Then, it is followed by the supply chain’s optimization problem. The government’s 440
decision-making problem in MS model is formulated as 441
max 𝑆𝑊 𝑀𝑆 (𝑠) = 𝜋𝑠𝑐

𝑀𝑆∗
+ 𝐶𝑆 𝑀𝑆∗ − 𝑠𝑞2𝑀𝑆∗ − 𝑡(𝑞1𝑀𝑆∗ + (1 − 𝑒)𝑞2𝑀𝑆∗ ) (12)
𝑠
the optimal subsidy by solving the government’s optimization problem. 443
Theorem 2. Under the differentiated subsidy strategy and MS model, the government’s optimal 444
subsidy that maximizes the total social welfare is 445
(𝜃 − 1 − 𝑐)(3𝑘(𝜃 − 1) + 𝜆(2𝑡 + 𝜆))

𝑠 𝐷𝑀𝑆∗ = (13)
𝑘(𝜃 − 1) − 2𝜆(𝑡 + 𝜆)
𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗

Corollary 4. < 0, < 0, > 0, > 0. 447
𝜕𝑘 𝜕𝑐 𝜕𝜆 𝜕𝑡
Similar to the C model, when the government adopts a differentiated subsidy strat- 448
egy, the increases of green product R&D cost coefficient and green product production 449
cost lead to the decrease of government subsidy. The government subsidy increases in 450
consumers’ green preference coefficient and negative environmental impact coefficient. 451
Substituting the optimal government subsidy given by Eq. (13) into the equilibrium 452
decision of the supply chain given by Lemma 3, we can derive Lemma 4 as follow. 453
Lemma 4. Under the differentiated subsidy strategy and MS model, the optimal wholesale prices, 454
retail prices and demands of the regular and green products, the optimal green product design level 455
1 𝑘(4𝜃−3−2𝑐)(𝜃−1)−2(1+2𝑐)𝜆(𝑡+𝜆) 3
are 𝑤1𝐷𝑀𝑆∗ = , 𝑤2𝐷𝑀𝑆∗ = , 𝑝1𝑀𝑆∗ = , 𝑝2𝐷𝑀𝑆∗ = 456
2 2𝑘(𝜃−1)−4𝜆(𝑡+𝜆) 4
𝑘(12𝜃−9−8𝑐)(𝜃−1)−2(3+4𝑐)𝜆(𝑡+𝜆) 1 𝑘(𝜃−1−𝑐) 𝑘(𝜃−1−𝑐)
, 𝑞1𝐷𝑀𝑆∗ = − , 𝑞2𝐷𝑀𝑆∗ = , 𝑒 𝐷𝑀𝑆∗ = 457
4𝑘(𝜃−1)−8𝜆(𝑡+𝜆) 4 𝑘(𝜃−1)−2𝜆(𝑡+𝜆) 𝑘(𝜃−1)−2𝜆(𝑡+𝜆)
(𝜃−1−𝑐)𝜆
. 458
𝑘(𝜃−1)−2𝜆(𝑡+𝜆)
4.3. Retailer-led Supply Chain (RS) Model 459

In this subsection, we consider the scenario where the retailer has more channel 460
power than the manufacturer. We call this scenario as retailer-led supply chain (RS) 461
model, in which the retailer acts as the Stackelberg game leader, while the manufacturer 462
acts as the follower. In the RS model, the retailer decides the retail prices of the regular 463
and green products. Then the manufacturer decides the green product design level and 464
the wholesale prices of the regular and green products. The RS model is formulated as 465
max 𝜋𝑟𝑅𝑆 (𝑝1 , 𝑝2 ) = (𝑝1 − 𝑤1∗ )𝑞1 + (𝑝2 − 𝑤2∗ )𝑞2

𝑝1 ,𝑝2
𝑤1∗ , 𝑤2∗ and 𝑒 ∗ are derived from solving the following problem, 466
𝑅𝑆 (𝑤
1 2
max 𝜋𝑚 1 , 𝑤2 , 𝑒) = 𝑤1 𝑞1 + (𝑤2 − 𝑐)𝑞2 − 𝑘𝑒
{ 𝑤1 ,𝑤2 ,𝑒 2
where 𝑞1 and 𝑞2 are given by Eqs. (2) and (3). Without loss of generality, let 𝑚1 and 467
𝑚2 denote profit margins of the regular and green products, respectively. Then, we have 468
(144
𝑝𝑖 = 𝑤𝑖 + 𝑚𝑖 , (𝑚𝑖 > 0, 𝑖 = 1,2).
)
Substituting Eq. (14) into the RS model, then we solve the RS model and Lemma 5 469
can be obtained. 470
Lemma 5. In the RS model, the optimal wholesale prices, retail prices and demands of the regular 471
1
and green products, and the optimal green product design level are 𝑤1𝑅𝑆∗ = , 𝑤2𝑅𝑆∗ = 472
4
2𝑘(𝜃−1)(3𝑐+𝑠+𝜃)−(1+4𝑐)𝜆2 3 2𝑘(𝜃−1)(𝑐+3(𝑠+𝜃))−𝜆2 −2(𝑐+𝑠+𝜃)𝜆2 2𝑘(𝑐−𝑠)−𝜆2
, 𝑝1∗ = , 𝑝2𝑅𝑆∗ = , 𝑞1𝑅𝑆∗ = , 473
8𝑘(𝜃−1)−4𝜆2 4 8𝑘(𝜃−1)−4𝜆2 8𝑘(𝜃−1)−4𝜆2
𝑘(𝑠+𝜃−1−𝑐) (𝑠+𝜃−1−𝑐)𝜆
𝑞2𝑅𝑆∗ = , 𝑒 𝑅𝑆∗
= . 474
4𝑘(𝜃−1)−2𝜆2 4𝑘(𝜃−1)−2𝜆2
According to Lemma 5, the following Corollary 5 can be obtained. 475
Corollary 5. In the RS model, 476
𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗

(1) for the regular product: > 0, > 0, < 0, < 0, < 0; 477
𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠
𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑝2𝑅𝑆∗ 𝜕𝑝2𝑅𝑆∗

(2) for the green product: < 0, > 0, > 0, > 0; < 0, > 478
𝜕𝑘 𝜕𝑐 𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐
𝜕𝑝2𝑅𝑆∗ 𝜕𝑝2𝑅𝑆∗ 𝜕𝑞2𝑅𝑆∗ 𝜕𝑞2𝑅𝑆∗ 𝜕𝑞2𝑅𝑆∗ 𝜕𝑞2𝑅𝑆∗ 𝜕𝑞2𝑅𝑆∗ 𝜕𝑒 𝑅𝑆∗ 𝜕𝑒 𝑅𝑆∗
0, > 0, > 0; < 0, < 0, > 0, > 0, > 0; < 0, < 479
𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐 𝜕𝜃 𝜕𝜆 𝜕𝑠 𝜕𝑘 𝜕𝑐
𝜕𝑒 𝑅𝑆∗ 𝜕𝑒 𝑅𝑆∗ 𝜕𝑒 𝑅𝑆∗
0, > 0, > 0, > 0. 480
𝜕𝜃 𝜕𝜆 𝜕𝑠
Similar to C and MS model, in the RS model, the wholesale price and retail price of 481
the regular product have nothing to do with model parameters. The green product R&D 482
cost coefficient has negative impact on the wholesale price and retail price of the green 483
product, while the green product production cost has positive impact on the wholesale 484
price and retail price of the green product. The green product R&D cost coefficient and 485
green product production cost both have negative impact on the market demand and 486
green design level of the green product, while they have positive impact on market de- 487
mand for the regular product. The consumers’ green preference and government subsidy 488
both have positive impact on the wholesale price, retail price, market demand and green 489
design level of the green product, while they have negative impact on the retail price of 490
the regular product. 491
In the following, we turn to the government’s problem in the RS model. Under the 492
differentiated subsidy strategy, we consider the government’s problem of maximizing en- 493
tire social welfare. The government first decides the optimal subsidy for the green prod- 494
uct. Then, it is followed by the supply chain’s optimization problem. The government’s 495
decision-making problem in RS model is formulated as 496
max 𝑆𝑊 𝑅𝑆 (𝑠) = 𝜋𝑠𝑐

𝑅𝑆∗
+ 𝐶𝑆 𝑅𝑆∗ − 𝑠𝑞2𝑅𝑆∗ − 𝑡(𝑞1𝑅𝑆∗ + (1 − 𝑒)𝑞2𝑅𝑆∗ ) (155
𝑠 )
the optimal subsidy by solving the government’s optimization problem. 498
Theorem 3. Under the differentiated subsidy strategy and RS model, the government’s optimal 499
subsidy that maximizes the total social welfare is 500
(𝜃 − 1 − 𝑐)(3𝑘(𝜃 − 1) + 2𝑡𝜆)
𝑠 𝐷𝑅𝑆∗ = (16)
𝑘(𝜃 − 1) − 2𝜆(𝑡 + 𝜆)
𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗

Corollary 6. < 0, < 0, > 0, > 0. 502
𝜕𝑘 𝜕𝑐 𝜕𝜆 𝜕𝑡
Similar to the C and MS model, the green product R&D cost coefficient and green 503
product production cost have negative impacts on the government subsidy, while the 504
consumers’ green preference and negative environmental impact coefficient have positive 505
impacts. 506
Substituting the optimal government subsidy given by Eq. (16) into the equilibrium 507
decision of the supply chain given by Lemma 5, we can derive Lemma 6 as follow. 508
Lemma 6. Under the differentiated subsidy strategy and RS model, the optimal wholesale prices, 509
retail prices and demands of the regular and green products, the optimal green product design level 510
1 𝑘(𝜃−1)(4𝜃−3)−2(1+4𝑐)𝜆(𝑡+𝜆) 3
are 𝑤1𝐷𝑅𝑆∗ = , 𝑤2𝐷𝑅𝑆∗ = , 𝑝1∗ = , 𝑝2𝐷𝑅𝑆∗ = 511
4 4𝑘(𝜃−1)−8𝜆(𝑡+𝜆) 4
𝑘(12𝜃−9−8𝑐)(𝜃−1)−2𝜆((3+4𝑐)𝑡+𝜆+2(𝑐+𝜃)𝜆) 1 𝑘(𝜃−1−𝑐) 𝑘(𝜃−1−𝑐)
, 𝑞1𝐷𝑅𝑆∗ = − , 𝑞2𝐷𝑅𝑆∗ = , 512
4𝑘(𝜃−1)−8𝜆(𝑡+𝜆) 4 𝑘(𝜃−1)−2𝜆(𝑡+𝜆) 𝑘(𝜃−1)−2𝜆(𝑡+𝜆)
(𝜃−1−𝑐)𝜆
𝑒 𝐷𝑅𝑆∗ = . 513
𝑘(𝜃−1)−2𝜆(𝑡+𝜆)
5. Comparisons and Managerial Implications 514

In this section, we compare the optimal decisions and profits of the supply chain and 515
the optimal subsidy given by the government under different channel power structures 516
(i.e., C, MS and RS models) and subsidy strategies of government. We also investigate the 517
impacts of key model parameters on the optimal decisions and profits (or social welfare) 518
of the supply chain and government through numerical analysis method. 519
5.1. Uniform Subsidy Strategy of Government 520

In this subsection, we consider the scenario where the government offers uniform 521
subsidies to the consumers regardless of the supply chain structures. Under this scenario, 522
we compare the optimal decisions in the C, MS and RS models and find some important 523
managerial implications. 524
Proposition 1. 𝑤1𝑀𝑆∗ = 2𝑤1𝑅𝑆∗ , 𝑝1𝑀𝑆∗ = 𝑝1𝑅𝑆∗ > 𝑝1𝐶∗ ; 𝑤2𝑀𝑆∗ > 𝑤2𝑅𝑆∗ , 𝑝2𝑀𝑆∗ > 𝑝2𝑅𝑆∗ > 𝑝2𝐶∗ . 525
The above Proposition compares the optimal wholesale prices and retailer prices of 526
the regular and green products under three models. From Proposition 1, due to the 527
manufacturer’s first-mover advantage, the wholesale prices of the regular and green prod- 528
ucts in the MS model are greater than those in the RS model. The retail price of the green 529
product in the MS model is greater than the retail price in the RS model, while the retail 530
price of the regular product in the MS model is equal to the retail price in the RS model. 531
Because of the double marginalization effect, the retail price of the regular and green prod- 532
ucts in MS and RS models are both greater than the retail price in the C model. 533
Proposition 2. 𝑞2𝐶∗ > 𝑞2𝑅𝑆∗ > 𝑞2𝑀𝑆∗ ; If 𝜆2 < 3𝑘(𝜃 − 1) − 𝑘√(𝜃 − 1)(9(𝜃 − 1) − 8(𝑐 − 𝑠)) , 534
𝑞 𝐶∗ 𝑞2𝑅𝑆∗ 𝑞2𝑀𝑆∗
then 𝑞1𝐶∗ > 𝑞1𝑀𝑆∗ > 𝑞1𝑅𝑆∗ ; otherwise, 𝑞1𝑀𝑆∗ > 𝑞1𝐶∗ > 𝑞1𝑅𝑆∗ ; 𝐶∗ 2 𝐶∗ = > . 535
𝑞1 +𝑞2 𝑞1𝑅𝑆∗ +𝑞2𝑅𝑆∗ 𝑞1𝑀𝑆∗ +𝑞2𝑀𝑆∗
Regarding the green product, the market demand for the green product in the RS 536
model is greater than that in the MS model, but both are less than the market demand in 537
the C model. It means that the retailer-led supply chain model can achieve wider market 538
coverage of the green product. Regarding the regular product, unlike the above scenario, 539
the demand for the regular product in the MS model is greater than that in the RS model. 540
It means that the manufacturing-led supply chain model can achieve wider market cov- 541
erage of the regular product. In summary, compared with the MS model, green product 542
can achieve a larger market coverage (i.e., 𝑞2𝑅𝑆∗ > 𝑞2𝑀𝑆∗ ) and a higher market share in the 543
𝑞2𝑅𝑆∗ 𝑞2𝑀𝑆∗
RS model (i.e., > ). Moreover, market share of the green product in the 544
𝑞1𝑅𝑆∗ +𝑞2𝑅𝑆∗ 𝑞1𝑀𝑆∗ +𝑞2𝑀𝑆∗
𝑞2𝐶∗ 𝑞2𝑅𝑆∗
RS model is same to that in the C model (i.e., = ). When consumers’ green 545
𝑞1𝐶∗ +𝑞2𝐶∗ 𝑞1𝑅𝑆∗ +𝑞2𝑅𝑆∗
preference coefficient is relatively small, the market demand for the regular product in the 546
MS model is less than that in the C model. Otherwise, the market demand for the regular 547
product in the MS model is more than that in the C model. The underlying managerial 548
implication is that establishing a retailer-led supply chain is more beneficial for the enter- 549
prise to expand the market share of the green products. 550
1
Proposition 3. 𝑒 𝐶∗ > 𝑒 𝑅𝑆∗ > 𝑒 𝑀𝑆∗ , 𝑒 𝑅𝑆∗ = 𝑒 𝐶∗ . 551
2
The green product design level in the RS model is greater than that in the MS model, 552
but it is smaller than that in the C model. From Propositions 1 and 2, compared with the 553
MS model, in the RS model, the green product can not only achieve a larger market cov- 554
erage, but its retail price is also lower. Moreover, the green product is greener and more 555
environmentally friendly in the RS model. Therefore, from the perspective of the green 556
design level, when the government provides a uniform subsidy strategy, it is a good 557
choice for the enterprise to develop a retailer-led supply chain structure. 558
5.2. Differentiated Subsidy Strategy of Government 559

In this subsection, we consider the scenario where the government offers differenti- 560
ated subsidies to the consumers based on different supply chain structures. We compare 561
the optimal decisions in three models and investigate the impacts of government subsidy 562
strategy on the supply chain decisions. Some important managerial implications are also 563
given. 564
Proposition 4. 𝑤1𝐷𝑀𝑆∗ = 2𝑤1𝐷𝑅𝑆∗ , 𝑝1𝐷𝑀𝑆∗ = 𝑝1𝐷𝑅𝑆∗ > 𝑝1𝐷𝐶∗ ; 𝑤2𝐷𝑀𝑆∗ > 𝑤2𝐷𝑅𝑆∗ , 𝑝2𝐷𝑀𝑆∗ > 𝑝2𝐷𝑅𝑆∗ . 565
According to Propositions 1 and 4, when the government provides a differentiated 566

subsidy strategy, the relationship between the wholesale price and retail price of the reg- 567
ular and green products in different supply chain models is same to the scenario when 568
the government provides a uniform subsidy strategy. For example, the wholesale and re- 569
tail prices of the green product in the MS model are higher than those in the RS model 570
under the differentiated subsidy strategy of government. This shows that the 571
government’s differentiated subsidy strategy doesn’t change the pricing relationships of 572
the regular and green products in C, MS and RS models. 573
𝑞2𝐷𝑅𝑆∗ 𝑞2𝐷𝑀𝑆∗
Proposition 5. 𝑞1𝐷𝐶∗ > 𝑞1𝐷𝑀𝑆∗ = 𝑞1𝐷𝑅𝑆∗ , 𝑞2𝐷𝑅𝑆∗ = 𝑞2𝐷𝑀𝑆∗ > 𝑞2𝐷𝐶∗ , 𝐷𝑅𝑆∗ = > 574
𝑞1 +𝑞2𝐷𝑅𝑆∗ 𝐷𝑀𝑆∗
𝑞1 +𝑞2𝐷𝑀𝑆∗
𝑞2𝐷𝐶∗
. 575
𝑞1𝐷𝐶∗ +𝑞2𝐷𝐶∗
Unlike the scenario where the government provides a uniform subsidy strategy, 576
when the government provides a differentiated subsidy strategy, the market demand of 577
the regular (green) product in MS model is same to that of regular (green) product in RS 578
model. That is to say, differentiated subsidy strategy of the government can eliminate the 579
impact of the different channel powers on the market demands for the regular and green 580
products. Combining Proposition 4, although the retail price of the green product under 581
different decentralized supply chain structures are different, the corresponding market 582
demands are indeed the same. Compared with the uniform subsidy strategy of govern- 583
ment, when the government provides a differentiated subsidy strategy, MS and RS supply 584
chain structures are equivalent. The enterprise can achieve more market demand (i.e., 585
𝑞2𝐷𝑅𝑆∗ 𝑞2𝐷𝑀𝑆∗ 𝑞2𝐷𝐶∗
𝑞2𝐷𝑅𝑆∗ = 𝑞2𝐷𝑀𝑆∗ > 𝑞2𝐷𝐶∗ ) and market share (i.e., = > ) for the 586
𝑞1𝐷𝑅𝑆∗ +𝑞2𝐷𝑅𝑆∗ 𝑞1𝐷𝑀𝑆∗ +𝑞2𝐷𝑀𝑆∗ 𝑞1𝐷𝐶∗ +𝑞2𝐷𝐶∗
green product in the MS and RS supply chain structures than in the C supply chain struc- 587
ture. The underlying managerial implication is that when the government provides a dif- 588
ferentiated subsidy strategy, establishing a decentralized supply chain structure can ena- 589
ble the enterprise to obtain a larger market coverage and market share. 590
Proposition 6. 𝑒 𝐷𝑅𝑆∗ = 𝑒 𝐷𝑀𝑆∗ > 𝑒 𝐷𝐶∗ . 591
When the government provides a differentiated subsidy strategy, the green product 592
design level in the RS model is same to that in the MS model. And, they are both greater 593
than the green product design level in the C model. Combining Proposition 3, the govern- 594
ment’s differentiated subsidy strategy can eliminate the influence of different supply 595
chain structures on green design level. From the perspective of product greenness degree, 596
the decentralized supply chain structure is better than the centralized supply chain struc- 597
ture under the government’s differentiated subsidy strategy. The underlying managerial 598
implication is that when the government provides a differentiated subsidy strategy, a de- 599
centralized supply chain is more conducive to the development of green products. 600
Proposition 7. 𝑠 𝐷𝑀𝑆∗ > 𝑠 𝐷𝑅𝑆∗ > 𝑠 𝐷𝐶∗ . 601
When the government provides a differentiated subsidy strategy, the government 602
subsidy level in the RS model is smaller than that in the MS model, but it is greater than 603
the subsidy level under the C model. That is to say, in order to maximize the total social 604
welfare, the RS model can enable the government to provide less subsidy compared with 605
the MS model. A higher subsidy from the government can make the MS supply chain 606
structure realize the same green design level with RS supply chain structure. 607
5.3. Numerical analysis 608

Due to the complexity, it is hard to compare the decision makers’ profits and total 609
social welfare under different supply chain structures and subsidy strategies of the gov- 610
ernment. Therefore, in this subsection we use numerical analysis method to obtain more 611
managerial implications. We assume that the model parameters satisfy: 𝑐 = 0.3, 𝜃 = 1.45, 612
𝑠 = 0.2, 𝑘 = 3, 𝑡 = 0.1. The model parameters are set based on the assumptions presented 613
in Section 3 and industry practice. This kind of simulation method is widely adopted by 614
the relevant literature (e.g., [5], [7], [13], [14] and [32]). The authors also conduct numerical 615
experiments with other combinations of model parameters, the results of which show that 616
the main conclusions and analytical results in this study are robust. Because consumers’ 617
environmental preference has important influence on the design of the green product and 618
supply chain structure, we investigate the impacts of consumers’ environmental prefer- 619
ence coefficient 𝜆 when 𝜆 is from 0.5 to 0.7. The above model parameters combination 620
satisfies the parameters constraints so that our models are solvable and our analysis is 621
effective. 622
5.3.1. Uniform subsidy strategy of government 623

Under the scenario where the government provides a uniform subsidy strategy, Figs. 624
1-3 intuitively demonstrate the impacts of consumers’ green preference coefficient on sup- 625
ply chain pricing decisions and supply chain profits under different supply chain struc- 626
tures. As shown in Fig. 1, the wholesale price of the green product in the MS model is 627
much greater than that in the RS model. The retail price of the green product in the MS 628
model is slightly higher than that in the RS model, while they are both greater than the 629
retail price in the C model. In addition, with the increase in the consumers’ green prefer- 630
ence, the wholesale price and retail price of the green product both increase in MS and RS 631
models, respectively. 632
633
Figure 1. 𝑤2𝑀𝑆∗ 𝑣𝑠. 𝑤2𝑅𝑆∗ , 𝑝2𝑀𝑆∗ 𝑣𝑠. 𝑝2𝑅𝑆∗ 634
Fig. 2 shows the changes in the market demands for the regular and green products 635
in the C, MS and RS models, respectively, when the consumers’ green preference coeffi- 636
cient varies. As the figure shows, the improvement of consumers’ green preference can 637
increase the market demand for the green product in any supply chain structures, while 638
reducing the market demand for the regular product in any supply chain structures. For 639
the green product, the market demand in the C model is the highest, followed by the RS 640
model and MS model. For the regular product, the market demand for the regular product 641
in the MS model is always greater than that in the RS model, and may be greater than the 642
market demand in the C model. From Fig. 2, we can also derive that green product can 643
achieve a larger market coverage and a higher market share in the RS model, compared 644
with the MS model. 645
646
Figure 2. 𝑞1𝐶∗ 𝑣𝑠. 𝑞1𝑀𝑆∗ 𝑣𝑠. 𝑞1𝑅𝑆∗ , 𝑞2𝐶∗ 𝑣𝑠. 𝑞2𝑀𝑆∗ 𝑣𝑠. 𝑞2𝑅𝑆∗ 647
Fig. 3 shows how the profits of the manufacturer and retailer in the MS and RS mod- 648
els vary with consumers’ green preference coefficient. As the figure shows, the profit of 649
the retailer in the RS model is greater than the profit of the manufacturer in the MS model. 650
The profit of the manufacturer in the RS model is also slightly greater than the profit of 651
the retailer in the MS model. This also shows the total profit of the supply chain in the RS 652
model is greater than that in the MS model. With the increase in consumers’ green prefer- 653
ence, the profits of the manufacturer and retailer in the two models both increase. 654
655
𝑀𝑆∗ 𝑅𝑆∗ ,
Figure 3. 𝜋𝑚 𝑣𝑠. 𝜋𝑚 𝜋𝑟𝑀𝑆∗ 𝑣𝑠. 𝜋𝑟𝑅𝑆∗ 656
Figs. 4 and 5 show how the green product design level and total social welfare change 657
with consumers’ green preference coefficient under different supply chain structures. Fig. 658
4 shows the green design level in the RS model is greater than that in the MS model, but 659
less than that in the C model. With the increase in consumers’ green preference, the green 660
design level in any supply chain models increases. Fig. 5 shows the total social welfare in 661
the RS model is slightly greater than that in the MS model, but less than the total social 662
welfare in the C model. With the increase in consumers’ green preference, the total social 663
welfare in any supply chain models rises. 664
665
𝐶∗ 𝑀𝑆∗ 𝑅𝑆∗
Figure 4. 𝑒 𝑣𝑠. 𝑒 𝑣𝑠. 𝑒 666
667
Figure 5. 𝑆𝑊 𝐶∗ 𝑣𝑠. 𝑆𝑊 𝑀𝑆∗ 𝑣𝑠. 𝑆𝑊 𝑅𝑆∗ 668
5.3.2. Differentiated subsidy strategy of government 669

Under the scenario where the government provides a differentiated subsidy strategy, 670
Figs. 6 and 7 visually demonstrate the impacts of consumers’ green preference coefficient 671
on supply chain pricing decisions and supply chain profits in different supply chain mod- 672
els. As Figs. 6 shows, the wholesale price of the green product in the MS model is greater 673
than that in the RS model. The retail price of the green product in the MS model is greater 674
than that in the RS model. In addition, with the increase in consumers’ green preference, 675
the wholesale price and retail price of the green product in the two supply chain models 676
both rise. 677
678
Figure 6. 𝑤2𝐷𝑀𝑆∗ 𝑣𝑠. 𝑤2𝐷𝑅𝑆∗ , 𝑝2𝐷𝑀𝑆∗ 𝑣𝑠. 𝑝2𝐷𝑅𝑆∗ 679
Fig. 7 shows how the profits of the manufacturer and retailer in the MS and RS mod- 680
els vary with consumers’ green preference coefficient. Contrary to the government’s uni- 681
form subsidy strategy, the profit of the manufacturer in the MS model is greater than the 682
profit of the retailer in the RS model. The profit of the retailer in the MS model is also 683
slightly greater than the profit of the manufacturer in the RS model. This also shows, when 684
the government offers a differential pricing strategy, the total profit of the supply chain in 685
the MS model is greater than that in the RS model. With the increase in consumers’ green 686
preference, the profits of the manufacturer and retailer in the MS and RS models both 687
increase. 688
689
𝐷𝑀𝑆∗ 𝐷𝑅𝑆∗ ,
Figure 7. 𝜋𝑚 𝑣𝑠. 𝜋𝑚 𝜋𝑟𝐷𝑀𝑆∗ 𝑣𝑠. 𝜋𝑟𝐷𝑅𝑆∗ 690
Fig. 8 shows how the green product design level varies with consumers’ green pref- 691
erence coefficient in different supply chain models. The green design level in the RS model 692
is equal to the green design level in the MS model, which is greater than that in the C 693
model, and is also greater than those in both MS and RS models under the scenario where 694
the government provides a uniform subsidy strategy (compared with Fig. 4). That is to 695
say, the government’s differentiated subsidy strategy can improve the green product 696
design level, and it can also eliminate the impacts of different supply chain structures on 697
the green design level. 698
699
Figure 8. 𝑒 𝐷𝐶∗ 𝑣𝑠. 𝑒 𝐷𝑀𝑆∗ 𝑣𝑠. 𝑒 𝐷𝑅𝑆∗ 700
Figs. 9 and 10 show how the government’s optimal subsidy and total social welfare 701
change with consumers’ green preference coefficient under different supply chain struc- 702
tures. As Fig. 9 shows, the government’s optimal subsidy in the RS model is less than that 703
in the MS model, but it is greater than that in the C model. As Fig. 10 shows, the total 704
social welfare in the RS model is equal to the total social welfare in the MS model, but it 705
is less than that in the C model. In conclusion, compared with the RS model, the govern- 706
ment needs to provide higher subsidy in the MS model to achieve the same total social 707
welfare. The government’s differentiated subsidy strategy can eliminate the impacts of 708
different supply chain structures on the total social welfare. 709
Different from the uniform subsidy strategy of government, with the increase in con- 710
sumers’ green preference, the total social welfare in the MS and RS models decreases. This 711
is because that the government needs to provide more subsidies to consumers in the MS 712
and RS models, while it reduces the total social welfare. In other words, as consumers’ 713
green preference increases, the expenditure of government subsidy exceeds the increase 714
in consumer utility and supply chain profit, as well as the improvement of the environ- 715
ment. Therefore, the total social welfare declines in consumers’ green preference. 716
717
Figure 9. 𝑠 𝐷𝐶∗ 𝑣𝑠. 𝑠 𝐷𝑀𝑆∗ 𝑣𝑠. 𝑠 𝐷𝑅𝑆∗ 718
719
Figure 10. 𝑆𝑊 𝐷𝐶∗ 𝑣𝑠. 𝑆𝑊 𝐷𝑀𝑆∗ 𝑣𝑠. 𝑆𝑊 𝐷𝑅𝑆∗ 720
6. Conclusions 721
To promote sustainable development in circular economy era, this paper studies the 722
multi-product pricing and green product design strategies in a sustainable supply chain 723
consisting of a manufacturer and a retailer considering the coexistence of the regular and 724
green products under different supply chain structures. We also investigate the impacts 725
of different government subsidy strategies, i.e., uniform subsidy strategy and differenti- 726
ated subsidy strategy, on the supply chain decisions and social welfare. By the utility 727
model, the consumer demand function is constructed. Considering different channel pow- 728
ers, we establish the C, MS and RS models, respectively. By solving the three models, the 729
optimal decisions of the supply chain and the government are obtained. In addition, un- 730
der both uniform subsidy strategy and differentiated subsidy strategy, we compare the 731
wholesale prices, retail prices, market demands, green product design level and the gov- 732
ernment’s optimal subsidies in different models. Finally, using numerical analysis 733
method, we discuss the impacts of the ever-changing operational environment on the 734
supply chain decisions, profits and total social welfare. Based on our research, we derive 735
the following conclusions and managerial implications. 736
1. The consumers’ green preference has positive impact on the design and development 737
of the green product and supply chain profits under different supply chain struc- 738
tures. However, the R&D cost and production cost have negative impact on the mar- 739
ket demand and green design level for the green product. Thus, it leads the firm re- 740
luctant to invest in the green product development. For the firm, in order to realize 741
the sustainable development of our society, it is beneficial to adopt various reasona- 742
ble measures to boost environmental awareness of consumers and decrease the R&D 743
and production costs of the green product. 744
2. When the government provides a uniform pricing strategy, in RS supply chain struc- 745
ture, the green product design level, total supply chain profit, and total social welfare 746
are all greater than those in the MS supply chain structure. Compared with the MS 747
supply chain structure, the green product can achieve a larger market coverage and 748
a higher market share in the RS supply chain structure. In other words, the retailer- 749
led supply chain structure is superior to the manufacturer-led supply chain structure. 750
From the perspective of the firm, establishing a retailer-led supply chain is beneficial 751
for the firm and can bring greener product, more market demands, more profit and 752
more social welfare. 753
3. When the government provides a differentiated subsidy strategy, MS and RS supply 754
chain structures can both bring greener product and more market demands than the 755
centralized supply chain. From the perspective of the firm, given the differentiated 756
subsidy strategy of the government, establishing a decentralized supply chain struc- 757
ture can enable the enterprise to obtain a larger market coverage and market share. 758
From the perspective of social welfare, MS and RS supply chain structures are equiv- 759
alent for the firm. From the perspective of supply chain profit, establishing a manu- 760
facturer-led supply chain structure can bring more profit for the firm. 761
4. For the government, compared with providing a uniform pricing strategy, providing 762
a differentiated subsidy strategy can alter the product market structure and the de- 763
velopment strategy of the green product. That is to say, the government’s differenti- 764
ated subsidy strategy can make the firm achieve larger market share for the green 765
product and more environmentally friendly and greener product. In addition, the 766
government’s differentiated subsidy strategy can eliminate the influence of the sup- 767
ply chain structure on the green product design, market demand and total social wel- 768
fare. 769
Our paper can guide the firm to establish appropriate supply chain structure and 770
implement green product development strategy to promote sustainable development. 771
However, there are also some limitations in our paper. For example, we consider two dif- 772
ferent consumer subsidies from government. In reality, the government could also pro- 773
vide subsidy to the manufacturer directly to promote the sustainable production. In the 774
future, we will consider the impacts of the manufacturer subsidy from government on the 775
green product design under different channel structures. In platform economy era, the 776
firm usually has multiple sales channels and faces retail competition. Thus, in the future, 777
it would be also interesting to explore the effects of retail competition on the green product 778
design. 779
Author Contributions: Conceptualization, K.X., G.S. and Y.W.; Funding acquisition, G.S., Y.W. and 780
S.N.; Methodology, K.X. and G.S.; Project administration, Y.W. and S.N.; Software, K.X. and G.S.; 781
Supervision, Y.W. and S.N.; Writing – original draft, K.X. and G.S.; Writing – review & editing, Y.W. 782
and S.N. 783
Funding: This research was funded by (i) National Natural Science Foundation of China, grant 784
number 71802143, 72002114; (ii) Humanities and Social Sciences Research Project of Ministry of Ed- 785
ucation of China, grant number 20YJC630178; (iii) Social Science Planning and Research Project of 786
Shandong Province, China, grant number 20DGLJ09. (iv) Fundamental Research Funds of Shan- 787
dong University, grant number 2020GN044. 788
Institutional Review Board Statement: Not applicable. 789
Informed Consent Statement: Not applicable. 790
Data Availability Statement: Not applicable. 791
Conflicts of Interest: The authors declare no conflict of interest. 792
Appendix A. 793
Proof of Lemma 1. 794

𝐶
Firstly, we verify whether 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) exists the optimal solutions. We have 795
2
𝜕 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) 2𝜃
2
=− < 0, 796
𝜕(𝑝1 ) 𝜃−1
𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒)
| 𝜕(𝑝1 )2 𝜕𝑝1 𝜕𝑝2 | = 4 > 0, 797
|𝜕 2 𝜋 (𝑝 , 𝑝 , 𝑒) 𝜕 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒)| 𝜃 − 1
2
𝑠𝑐 1 2
𝜕𝑝1 𝜕𝑝2 𝜕(𝑝2 )2
𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒)
| 𝜕(𝑝1 )2 𝜕𝑝1 𝜕𝑝2 𝜕𝑝1 𝜕𝑒 |
𝜕 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒)
2 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) −4𝑘(𝜃 − 1) + 2𝜆2
𝐻(𝜋𝑠𝑐 ) = = . 798
𝜕𝑝2 𝜕𝑝1 𝜕(𝑝2 )2 𝜕𝑝2 𝜕𝑒 (𝜃 − 1)2
|𝜕 2 𝜋 (𝑝 , 𝑝 , 𝑒) 𝜕 2 𝜋 (𝑝 , 𝑝 , 𝑒) 𝜕 2 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒)|
𝑠𝑐 1 2 𝑠𝑐 1 2
𝜕𝑒𝜕𝑝1 𝜕𝑒𝜕𝑝2 𝜕(𝑒)2
Because of 𝑘(𝜃 − 1) > 2𝜆(2𝑡 + 𝜆) , then 4𝑘(𝜃 − 1) > 2𝜆2 and 𝐻(𝜋𝑠𝑐 ) < 0 . 799
Hence, 𝐻(𝜋𝑠𝑐 ) is negative definite. The objective function 𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) is jointly concave 800
in 𝑝1 , 𝑝2 and 𝑒 . By the first order conditions (FOC), let the first derivatives of 801
𝜋𝑠𝑐 (𝑝1 , 𝑝2 , 𝑒) with respect to (w.r.t.) 𝑝1 , 𝑝2 and 𝑒 equal to 0. We derive the optimal solu- 802
tions satisfying 803
1
𝑝1𝐶∗ =
2
2𝑘(𝜃 − 1)(𝑐 + 𝑠 + 𝜃) − (1 + 2𝑐)𝜆2
𝑝2𝐶∗ = . 804
4𝑘(𝜃 − 1) − 2𝜆2
(𝑠 + 𝜃 − 𝑐 − 1)𝜆
𝑒 𝐶∗ =
{ 2𝑘(𝜃 − 1) − 𝜆2
Therefore, in the C model, the optimal demands 𝑞1𝐶∗ and 𝑞2𝐶∗ of the regular and 805
2𝑘(𝑐−𝑠)−𝜆2 𝑘(𝑠+𝜃−𝑐−1)
green products are 𝑞1𝐶∗ = , 𝑞2𝐶∗ = . 806
4𝑘(𝜃−1)−2𝜆2 2𝑘(𝜃−1)−𝜆2
Then, the Lemma 1 is proved. □ 807
Proof of Corollary 1: 808
Based on the assumption conditions of 2𝑘(𝑐 − 𝑠) > 𝜆2 , 𝜃 − 1 > 𝑐 and 𝑘(𝜃 − 1) > 809
2𝜆(2𝑡 + 𝜆), it is easy to derive that 𝑠 + 𝜃 − 1 − 𝑐 > 0 and 𝑘(𝜃 − 1) > 𝜆2 . Therefore, (1) 810
𝜕𝑞1𝐶∗ (𝑠+𝜃−1−𝑐)𝜆2 𝜕𝑞1𝐶∗ 2𝑘 𝜕𝑞1𝐶∗ 𝑘(2𝑘(𝑐−𝑠)−𝜆2 ) 𝜕𝑞1𝐶∗
= > 0, = > 0, = − (2𝑘(𝜃−1)−𝜆2 )2 < 0, = 811
𝜕𝑘 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝑐 4𝑘(𝜃−1)−2𝜆2 𝜕𝜃 𝜕𝜆
2𝑘(𝑠+𝜃−1−𝑐)𝜆 𝜕𝑞1𝐶∗ 𝑘 𝜕𝑝2𝐶∗ (𝜃−1)(𝑠+𝜃−1−𝑐)𝜆2 𝜕𝑝2𝐶∗
− (2𝑘(𝜃−1)−𝜆2 )2 < 0, =− <0 ; (2) =− < 0, = 812
𝜕𝑠 2𝑘(𝜃−1)−𝜆2 𝜕𝑘 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝑐
2𝑘(𝜃−1)−2𝜆2 𝜕𝑝2𝐶∗ 𝑘(2𝑘(𝜃−1)2 +(2+𝑐−𝑠−2𝜃)𝜆2 ) 𝜕𝑝2𝐶∗ 2𝑘(𝑠+𝜃−1−𝑐)(𝜃−1)𝜆 𝜕𝑝2𝐶∗
> 0, = > 0, = > 0, = 813
4𝑘(𝜃−1)−2𝜆2 𝜕𝜃 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝜆 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝑠
𝑘(𝜃−1) 𝜕𝑞2𝐶∗ (𝑠+𝜃−1−𝑐)𝜆2 𝜕𝑞2𝐶∗ 𝑘 𝜕𝑞2𝐶∗ 𝑘(2𝑘(𝑐−𝑠)−𝜆2 )
>0 ; = − (2𝑘(𝜃−1)−𝜆2 )2 < 0, =− < 0, = (2𝑘(𝜃−1)−𝜆2 )2 > 814
2𝑘(𝜃−1)−𝜆2 𝜕𝑘 𝜕𝑐 2𝑘(𝜃−1)−𝜆2 𝜕𝜃
𝜕𝑞2𝐶∗ 2𝑘(𝑠+𝜃−1−𝑐)𝜆 𝜕𝑞2𝐶∗ 𝑘 𝜕𝑒 𝐶∗ 2(𝑠+𝜃−1−𝑐)(𝜃−1)𝜆 𝜕𝑒 𝐶∗

0, = (2𝑘(𝜃−1)−𝜆2 )2 > 0, = >0 ; =− < 0, = 815
𝜕𝜆 𝜕𝑠 2𝑘(𝜃−1)−𝜆2 𝜕𝑘 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝑐
𝜆 𝜕𝑒 𝐶∗ 𝜆(2𝑘(𝑐−𝑠)−𝜆2 ) 𝜕𝑒 𝐶∗ (𝑠+𝜃−1−𝑐)(2𝑘(𝜃−1)+𝜆2 ) 𝜕𝑒 𝐶∗ 𝜆
− < 0, = (2𝑘(𝜃−1)−𝜆2 )2 > 0, = > 0, = > 816
2𝑘(𝜃−1)−𝜆2 𝜕𝜃 𝜕𝜆 (2𝑘(𝜃−1)−𝜆2 )2 𝜕𝑠 2𝑘(𝜃−1)−𝜆2
0. 817
Then, the Corollary 1 is proved. □ 818
Proof of Theorem 1: 819
Substituting the optimal decisions given by Lemma 1 into the government’s objective 820
function given by Eq. (10), we have 821
1
𝑆𝑊 𝐶 (𝑠) = (4𝑘 2 (𝜃 − 1)(3𝑐 2 − 𝑠 2 + 2𝑠(𝜃 − 1) − (4𝑡 − 3𝜃)(𝜃 − 1) 822
8(2𝑘(𝜃 − 1) − 𝜆2 )2
− 2𝑐(𝑠 + 3𝜃 − 3)) + 8𝑘𝑡(𝑠 + 𝜃 − 𝑐 − 1)2 𝜆 823
− 4𝑘(−2 + 𝑐(2 + 𝑐) − 𝑠 2 + 4𝑡 + 𝜃 − 2(𝑐 + 2𝑡)𝜃 + 𝜃 2 )𝜆2 + (3 − 4𝑡)𝜆4 ). 824
Taking the second order derivative of 𝑆𝑊 𝐶 (𝑠) w.r.t. 𝑠, we have 825

𝐶
𝑑𝑆𝑊 (𝑠) 𝑘(𝜆(2𝑡 + 𝜆) − 𝑘(𝜃 − 1))
= . 826
𝑑𝑠 (2𝑘(𝜃 − 1) − 𝜆2 )2
𝑑𝑆𝑊 𝐶 (𝑠)
Because of 𝑘(𝜃 − 1) > 2𝜆(2𝑡 + 𝜆) , then 𝑘(𝜃 − 1) > 𝜆(2𝑡 + 𝜆) , and <0. 827
𝑑𝑠
𝐶
Hence, 𝑆𝑊 (𝑠) has the optimal solution. By the FOC, the optimal subsidy of government 828
is 829
(𝜃 − 1 − 𝑐)(𝑘(𝜃 − 1) + 2𝑡𝜆)
𝑠 𝐶∗ = . 830
𝑘(𝜃 − 1) − 𝜆(2𝑡 + 𝜆)
Then, the Theorem 1 is proved. □ 831
Proof of Corollary 2: 832
Similar to the proof of Corollary 1, it is omitted here because of the simplicity. □ 833
Proof of Theorem 2: 834
Substituting the optimal subsidy of government given by Theorem 1 into the equi- 835
librium decisions of the supply chain given by Lemma 1, the Theorem 2 can be derived. 836
Proof of Lemma 2: 837
Firstly, we solve the retailer’s optimization problem. We verify whether 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) 838
exists the optimal solutions. Because of 839
𝜕 2 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) 2𝜃
2
=− < 0, 840
𝜕(𝑝1 ) 𝜃−1
𝜕 2 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) 𝜕 2 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 )
| 𝜕(𝑝1 )2 𝜕𝑝1 𝜕𝑝2 | 4
|𝜕 2 𝜋 𝑀𝑆 (𝑝 , 𝑝 ) 2 𝑀𝑆 | = > 0, 841
𝑟 1 2 𝜕 𝜋𝑟 (𝑝1 , 𝑝2 ) 𝜃−1
𝜕𝑝1 𝜕𝑝2 𝜕(𝑝2 )2
hence 𝐻(𝜋𝑟𝑀𝑆 ) is negative definite. 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) is jointly concave in 𝑝1 and 𝑝2 . By the 842
FOC, the optimal pricing decisions of retailer are 843
1 + 𝑤1
𝑝1𝑀𝑆∗ =
{ 2 . 844
𝑠 + 𝜃 + 𝜆𝑒 + 𝑤2
𝑝2𝑀𝑆∗ =
2
Secondly, we solve the manufacturer’s optimization problem. We verify whether 845

𝑀𝑆
𝜋𝑚 (𝑤1 , 𝑤2 , 𝑒) exists the optimal solutions. We have 846
𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜃
=− < 0, 847
𝜕(𝑤1 )2 𝜃−1
𝜕 2 𝜋𝑚𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒)
| 𝜕(𝑤1 )2 𝜕𝑤1 𝜕𝑤2 | = 1 > 0, 848
|𝜕 2 𝜋 𝑀𝑆 (𝑤 , 𝑤 , 𝑒) 𝜕 𝜋𝑚 (𝑤1 , 𝑤2 , 𝑒)| 𝜃 − 1
2 𝑀𝑆
𝑚 1 2
𝜕𝑤1 𝜕𝑤2 𝜕(𝑤2 )2
𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒)
𝜕(𝑤 2
| 1) 𝜕𝑤1 𝜕𝑤2 𝜕𝑤1 𝜕𝑒 |
2 𝑀𝑆
𝜕 𝜋𝑚 (𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) −4𝑘(𝜃 − 1) + 𝜆2
𝐻(𝜋𝑚 ) = = . 849
𝜕𝑤2 𝜕𝑤1 𝜕(𝑤2 )2 𝜕𝑤2 𝜕𝑒 4(𝜃 − 1)2
|𝜕 2 𝜋 𝑀𝑆 (𝑤 , 𝑤 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒) 𝜕 2 𝜋𝑚
𝑀𝑆
(𝑤1 , 𝑤2 , 𝑒)|
𝑚 1 2
𝜕𝑒𝜕𝑤1 𝜕𝑒𝜕𝑤2 𝜕(𝑒)2
Because of 2𝑘(𝑐 − 𝑠) > 𝜆2 and 𝜃 − 1 > 𝑐 − 𝑠, then 4𝑘(𝜃 − 1) > 𝜆2 and 𝐻(𝜋𝑚 𝑀𝑆 )
< 850
𝑀𝑆 ) 𝑀𝑆
0. Hence, 𝐻(𝜋𝑚 is negative definite. 𝜋𝑚 (𝑤1 , 𝑤2 , 𝑒) is jointly concave in 𝑤1 , 𝑤2 and 𝑒. 851
By the FOC, the optimal decisions of the manufacturer are 852
1
𝑤1𝑀𝑆∗ =
2
4𝑘(𝜃 − 1)(𝑐 + 𝑠 + 𝜃) − (1 + 2𝑐)𝜆2
𝑤2𝑀𝑆∗ = . 853
8𝑘(𝜃 − 1) − 2𝜆2
(𝑠 + 𝜃 − 1 − 𝑐)𝜆
𝑒 𝑀𝑆∗ = .
{ 4𝑘(1 − 𝜃) − 𝜆2
Substituting the above optimal decisions into the optimal prices of retailer, we can 854
1
derive the equilibrium decisions of the supply chain as follows: 𝑤1𝑀𝑆∗ = , 𝑤2𝑀𝑆∗ = 855
2
4𝑘(𝜃−1)(𝑐+𝑠+𝜃)−(1+2𝑐)𝜆2 (𝑠+𝜃−1−𝑐)𝜆 4𝑘(𝑐−𝑠)−𝜆2 𝑘(𝑠+𝜃−1−𝑐) 3
, 𝑒 𝑀𝑆∗ = , 𝑞1𝑀𝑆∗ = , 𝑞2𝑀𝑆∗ = , 𝑝1𝑀𝑆∗ = , 856
8𝑘(𝜃−1)−2𝜆2 4𝑘(1−𝜃)−𝜆2 4(4𝑘(𝜃−1)−𝜆2 ) 4𝑘(𝜃−1)−𝜆2 4
4𝑘(𝜃−1)(𝑐+3(𝑠+𝜃))−(3+4𝑐)𝜆 2
𝑝2𝑀𝑆∗ = . 857
16𝑘(𝜃−1)−4𝜆2
Then, the Lemma 2 is proved. □ 858
Similar to the previous proofs, the proofs of Lemmas 3, Theorems 3-6, Corollaries 2- 859
6 are omitted. □ 860
Proof of Proposition 1: 861
Based on the assumption conditions of 2𝑘(𝑐 − 𝑠) > 𝜆2 , 𝜃 − 1 > 𝑐 and 𝑘(𝜃 − 1) > 862
2𝜆(2𝑡 + 𝜆), we can derive 𝑠 + 𝜃 − 1 − 𝑐 > 0 and 𝑘(𝜃 − 1) > 𝜆2 . Therefore,𝑤2MS∗ − 𝑤2RS∗ = 863
2𝑘(𝑠+𝜃−𝑐)(𝜃−1)(4𝑘(𝜃−1)−3𝜆2 )+𝜆4 (𝑠+𝜃−1−𝑐)𝜆2 (𝑘(𝜃−1)−𝜆2 )
> 0, 𝑝2MS∗ − 𝑝2RS∗ = > 0, 𝑝2RS∗ − 𝑝2C∗ = 864
4(4𝑘(𝜃−1)−𝜆2 )(2𝑘(𝜃−1)−𝜆2 ) 2(4𝑘(𝜃−1)−𝜆2 )(2𝑘(𝜃−1)−𝜆2 )
𝜆2 +2(𝑠+𝜃−𝑐)(𝑘(𝜃−1)−𝜆2 ) 2𝑘(𝜃−1)−𝜆2
> = 0. 865
8𝑘(𝜃−1)−4𝜆2 8𝑘(𝜃−1)−4𝜆2
Then, the Proposition 1 is proved. □ 866
Proof of Proposition 2: 867
Based on the previous proofs, we have the conditions of 𝑠 + 𝜃 − 1 − 𝑐 > 0 and 868
𝑘(𝑠+𝜃−1−𝑐)𝜆2
2
𝑘(𝜃 − 1) > 𝜆 . Therefore, 𝑞2𝑀𝑆∗ − 𝑞2𝑅𝑆∗ =− <0 , 𝑞2𝐶∗ − 𝑞2𝑅𝑆∗ = 869
2(4𝑘(𝜃−1)−𝜆2 )(2𝑘(𝜃−1)−𝜆2 )
𝑘(𝑠+𝜃−1−𝑐) 𝑘(𝑠+𝜃−1−𝑐)𝜆2 −2𝑐𝑘+2𝑘𝑠+𝜆2
> 0 ; 𝑞1𝑀𝑆∗ − 𝑞1𝑅𝑆∗ = > 0 , 𝑞1𝐶∗ − 𝑞1𝑅𝑆∗ = = 870
4𝑘(𝜃−1)−2𝜆2 2(4𝑘(𝜃−1)−𝜆2 )(2𝑘(𝜃−1)−𝜆2 ) −8𝑘(𝜃−1)+4𝜆2
2𝑘(𝑐−𝑠)−𝜆2 𝑞2𝐶∗ 𝑞2𝑅𝑆∗ 𝑞2𝑅𝑆∗ 𝑞2𝑀𝑆∗ 2𝑘(𝑠+𝜃−1−𝑐)𝜆2
> 0, − = 0, − = > 0. 871
8𝑘(𝜃−1)−4𝜆2 𝑞1𝐶∗ +𝑞2𝐶∗ 𝑞1𝑅𝑆∗ +𝑞2𝑅𝑆∗ 𝑞1𝑅𝑆∗ +𝑞2𝑅𝑆∗ 𝑞1𝑀𝑆∗ +𝑞2𝑀𝑆∗ 8𝑘 2 (𝜃−1)2 −6𝑘(𝜃−1)𝜆2 +𝜆4
2
(3𝑘(𝜃−1)−𝜆2 ) −𝑘 2 (𝜃−1)(9(𝜃−1)−8(𝑐−𝑠))
Because of 𝑞1𝐶∗ − 𝑞1𝑀𝑆∗ = , (1) If (3𝑘(𝜃 − 1) − 𝜆2 )2 > 872
4(4𝑘(𝜃−1)−𝜆2 )(2𝑘(𝜃−1)−𝜆2 )
𝑘 2 (𝜃 − 1)(9(𝜃 − 1) − 8(𝑐 − 𝑠)), that is 𝜆 < 3𝑘(𝜃 − 1) − 𝑘√(𝜃 − 1)(9(𝜃 − 1) − 8(𝑐 − 𝑠)),
2
873
then 𝑞1𝐶∗ − 𝑞1𝑀𝑆∗ > 0; (2) otherwise, 𝑞1𝐶∗ − 𝑞1𝑀𝑆∗ < 0. 874
Then, the Proposition 2 is proved. 875
Similar to the proofs of Propositions 1 and 2, we omit the proofs of Propositions 3-7. 876
877
References 878
1. Wu, Z., Pagell, M. Balancing priorities: Decision-making in sustainable supply chain management. J. Oper. Manag. 2011, 29, 577- 879
590. 880
2. Linton, J. D., Klassen, R., Jayaraman, V. Sustainable supply chains: An introduction. J. Oper. Manag. 2007, 25, 1075-1082. 881
3. Li, G., Li, L., Choi, T. M., Sethi, S. P. Green supply chain management in Chinese firms: Innovative measures and the moderating 882
role of quick response technology. J. Oper. Manag. 2020, 66, 958-988. 883
4. Shamdasani, P., Chon-Lin, G. O., Richmond, D. Exploring green consumers in an oriental culture: Role of personal and market- 884
ing mix factors. NA-Adv. Consum. Res. 1993, 20, 488-493. 885
5. Agi, M. A., Yan, X. Greening products in a supply chain under market segmentation and different channel power structures. 886
Int. J. Prod. Econ. 2020, 223, 107523. 887
6. Hong, Z., Wang, H., Gong, Y. Green product design considering functional-product reference. Int. J. Prod. Econ. 2019, 210, 155- 888
168. 889
7. Bian, J., Zhang, G., Zhou, G. Manufacturer vs. consumer subsidy with green technology investment and environmental concern. 890
Eur. J. Oper. Res. 2020, 287, 832-843. 891
8. Su, J. C., Wang, L., Ho, J. C. The impacts of technology evolution on market structure for green products. Math. Comput. Model. 892
2012, 55, 1381-1400. 893
9. Li, X., Li, Y. On green market segmentation under subsidy regulation. Supply Chain Manag. 2017, 22, 284-294. 894
10. Li, X., Li, Y. Chain-to-chain competition on product sustainability. J. Clean. Prod. 2016, 112, 2058-2065. 895
11. Sarkar, B., Mridha, B., Pareek, S., Sarkar, M., & Thangavelu, L. A flexible biofuel and bioenergy production system with trans- 896
portation disruption under a sustainable supply chain network. J. Clean. Prod. 2021, 317, 128079. 897
12. Deng, W., Feng, L., Zhao, X., Lou, Y. Effects of supply chain competition on firms’ product sustainability strategy. J. Clean. Prod. 898
2020, 275, 124061. 899
13. Sim, J., El Ouardighi, F., Kim, B. Economic and environmental impacts of vertical and horizontal competition and integration. 900
Nav. Res. Log. 2019, 66, 133-153. 901
14. Gao, J., Xiao, Z., Wei, H. Competition and coordination in a dual-channel green supply chain with an eco-label policy. Comput. 902
Ind. Eng. 2021, 153, 107057. 903
15. Zhang, L., Wang, J., You, J. Consumer environmental awareness and channel coordination with two substitutable products. 904
Eur. J. Oper. Res. 2015, 241, 63-73. 905
16. Hong, Z., Wang, H., Yu, Y. Green product pricing with non-green product reference. Transport. Res. E-Log. 2018, 115, 1-15. 906
17. Zhang, Q., Zhao, Q., Zhao, X., Tang, L. On the introduction of green product to a market with environmentally conscious con- 907
sumers. Comput. Ind. Eng. 2020, 139, 106190. 908
18. Yadav, D., Kumari, R., Kumar, N., & Sarkar, B. Reduction of waste and carbon emission through the selection of items with 909
cross-price elasticity of demand to form a sustainable supply chain with preservation technology. J. Clean. Prod. 2021, 297, 910
126298. 911
19. Ghosh, D., Shah, J. Supply chain analysis under green sensitive consumer demand and cost sharing contract. Int. J. Prod. Econ. 912
2015, 164, 319-329. 913
20. Hong, Z., Guo, X. Green product supply chain contracts considering environmental responsibilities. Omega, 2019, 83, 155-166. 914
21. Dong, C., Liu, Q., Shen, B. To be or not to be green? Strategic investment for green product development in a supply chain. 915
Transport. Res. E-Log. 2019, 131, 193-227. 916
22. Li, G., Wu, H., Sethi, S. P., Zhang, X. Contracting green product supply chains considering marketing efforts in the circular 917
economy era. Int. J. Prod. Econ. 2021, 234, 108041. 918
23. Zhu, W., He, Y. Green product design in supply chains under competition. Eur. J. Oper. Res. 2017, 258, 165-180. 919
24. Guo, S., Choi, T. M., Shen, B. Green product development under competition: A study of the fashion apparel industry. Eur. J. 920
Oper. Res. 2020, 280, 523-538. 921
25. Ma, P., Zhang, C., Hong, X., Xu, H. Pricing decisions for substitutable products with green manufacturing in a competitive 922
supply chain. J. Clean. Prod. 2018, 183, 618-640. 923
26. Du, P., Yang, X., Xu, L., Tan, Y., Li, H. Green design strategies of competing manufacturers in a sustainable supply chain. J. 924
Clean. Prod. 2020, 265, 121853. 925
27. Shan, H., Zhang, C., Wei, G. Bundling or Unbundling? Pricing Strategy for Complementary Products in a Green Supply Chain. 926
Sustainability, 2020, 12, 1331. 927
28. Cao, Y., Tao, L., Wu, K., Wan, G. Coordinating joint greening efforts in an agri-food supply chain with environmentally sensitive 928
demand. J. Clean. Prod. 2020, 277, 123883. 929
29. Guo, X., Cheng, L., Liu, J. Green supply chain contracts with eco-labels issued by the sales platform: profitability and environ- 930
mental implications. Int. J. Prod. Res. 2020, 58, 1485-1504. 931
30. Wang, Z., Zhang, J., Zhao, H. The Selection of Green Technology Innovations under Dual-Credit Policy. Sustainability, 2020, 12, 932
6343. 933
31. Yang, X., Xu, M., Zhang, W. Can Design for the Environment be Worthwhile? Green Design for Manufacturers Brands When 934
Confronted with Competition from Store Brands. Sustainability, 2020, 12, 1078. 935
32. Cai, Y., Choi, T., Feng, L., Li, Y. Producer’s Choice of Design-for-Environment under Environmental Taxation. Eur. J. Oper. Res. 936
2021. https://doi.org/10.1016/j.ejor.2021.04.048. 937
33. Mu, Z., Zheng, Y., Sun, H. Cooperative Green Technology Innovation of an E-Commerce Sales Channel in a Two-Stage Supply 938
Chain. Sustainability, 2021, 13, 7499. 939
34. Garai, A., Chowdhury, S., Sarkar, B., & Roy, T. K. Cost-effective subsidy policy for growers and biofuels-plants in closed-loop 940
supply chain of herbs and herbal medicines: An interactive bi-objective optimization in T-environment. Appl. Soft. Comput. 2021, 941
100, 106949. 942
35. Yalabik, B., Fairchild, R. J. Customer, regulatory, and competitive pressure as drivers of environmental innovation. Int. J. Prod. 943
Econ. 2011, 131, 519-527. 944
36. Subramanian, R., Gupta, S., Talbot, B. Product design and supply chain coordination under extended producer responsibility. 945
Prod. Oper. Manag. 2009, 18, 259-277. 946
37. Shen, B., Zhu, C., Li, Q., Wang, X. Green technology adoption in textiles and apparel supply chains with environmental taxes. 947
Int. J. Prod. Res. 2020, 1-18. 948
38. Chen, J. Y., Dimitrov, S., Pun, H. The impact of government subsidy on supply Chains’ sustainability innovation. Omega. 2019, 949
86, 42-58. 950
39. Atasu, A., Subramanian, R. Extended producer responsibility for e-waste: Individual or collective producer responsibility?. 951
Prod. Oper. Manag. 2012, 21, 1042-1059. 952
40. Guo, S., Choi, T. M., Shen, B. Green product development under competition: A study of the fashion apparel industry. Eur. J. 953
Oper. Res. 2020, 280, 523-538. 954
41. Carus, M., Eder, A., Beckmann, J. GreenPremium prices along the value chain of biobased products. Ind. Biotechnol. 2014, 10, 955
83-88. 956
42. Zhang, C. T., Wang, H. X., Ren, M. L. Research on pricing and coordination strategy of green supply chain under hybrid pro- 957
duction mode. Comput. Ind. Eng. 2014, 72, 24-31. 958
43. Ferrer, G., Swaminathan, J. M. Managing new and remanufactured products. Manag. Sci. 2006, 52, 15-26. 959
44. Yenipazarli, A. Managing new and remanufactured products to mitigate environmental damage under emissions regulation. 960
Eur. J. Oper. Res. 2016, 249, 117-130. 961
45. Chen, C. Design for the environment: A quality-based model for green product development. Manag. Sci. 2001, 47, 250-263. 962
46. Yalabik, B., Fairchild, R. J. Customer, regulatory, and competitive pressure as drivers of environmental innovation. Int. J. Prod. 963
Econ. 2011, 131, 519-527. 964
47. Gao, J., Xiao, Z., Wei, H., Zhou, G. Active or passive? Sustainable manufacturing in the direct-channel green supply chain: A 965
perspective of two types of green product designs. Transport. Res. D-Tr. E. 2018, 65, 332-354. 966
967

Sustainability 1406486

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sustainability 1406486

Uploaded by

Copyright:

Available Formats

Article 1

Optimal Pricing and Green Product Design Strategies in a Sus- 2

tainable Supply Chain Considering Government Subsidy and 3

Different Channel Power Structures 4

Sustainability 2021, 13, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/sustainability

1 Available at: https://newsroom.accenture.com/article_display.cfm?article_id=4801, accessed on 11 July 2021.

4 Available at: http://www.ujala.gov.in/, accessed on 11 July 2021.

2. Literature Review 143

2.1. Multi-product Pricing 148

2.2. Green Product Design 186

2.3. Government Policy 216

2.4. Summary 238

Issues involved Decision variables

3. Model Descriptions 261

5 Available at: http://www.triplepundit.com/2014/11/economics-sustainable-coffee-production/, accessed on 11 July,

Table 2. The notations used in this paper. 320

𝑆𝑊 = 𝜋𝑠𝑐 + 𝐶𝑆 − 𝑠𝑞2 − 𝑡(𝑞1 + (1 − 𝑒)𝑞2 ). (7)

4. Model Analysis 337

4.1. Centralized Supply Chain (C) Model 345

4.2. Manufacturer-led Supply Chain (MS) Model 403

max 𝜋𝑟𝑀𝑆 (𝑝1 , 𝑝2 ) = (𝑝1 − 𝑤1 )𝑞1 + (𝑝2 − 𝑤2 )𝑞2

4𝑘(𝜃−1)(𝑐+𝑠+𝜃)−(1+2𝑐)𝜆2 3 4𝑘(𝜃−1)(𝑐+3(𝑠+𝜃))−(3+4𝑐)𝜆2 4𝑘(𝑐−𝑠)−𝜆2

According to Lemma 3, the following Corollary 3 can be obtained. 418

Corollary 3. In the MS model, 419

𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗ 𝜕𝑞1𝑀𝑆∗

𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑤2𝑀𝑆∗ 𝜕𝑝2𝑀𝑆∗ 𝜕𝑝2𝑀𝑆∗

max 𝑆𝑊 𝑀𝑆 (𝑠) = 𝜋𝑠𝑐

(𝜃 − 1 − 𝑐)(3𝑘(𝜃 − 1) + 𝜆(2𝑡 + 𝜆))

𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗ 𝜕𝑠 𝐷𝑀𝑆∗

4.3. Retailer-led Supply Chain (RS) Model 459

max 𝜋𝑟𝑅𝑆 (𝑝1 , 𝑝2 ) = (𝑝1 − 𝑤1∗ )𝑞1 + (𝑝2 − 𝑤2∗ )𝑞2

According to Lemma 5, the following Corollary 5 can be obtained. 475

Corollary 5. In the RS model, 476

𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗ 𝜕𝑞1𝑅𝑆∗

𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑤2𝑅𝑆∗ 𝜕𝑝2𝑅𝑆∗ 𝜕𝑝2𝑅𝑆∗

max 𝑆𝑊 𝑅𝑆 (𝑠) = 𝜋𝑠𝑐

𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗ 𝜕𝑠 𝐷𝑅𝑆∗

5. Comparisons and Managerial Implications 514

5.1. Uniform Subsidy Strategy of Government 520

5.2. Differentiated Subsidy Strategy of Government 559

According to Propositions 1 and 4, when the government provides a differentiated 566

Proposition 6. 𝑒 𝐷𝑅𝑆∗ = 𝑒 𝐷𝑀𝑆∗ > 𝑒 𝐷𝐶∗ . 591

Proposition 7. 𝑠 𝐷𝑀𝑆∗ > 𝑠 𝐷𝑅𝑆∗ > 𝑠 𝐷𝐶∗ . 601

5.3. Numerical analysis 608

5.3.1. Uniform subsidy strategy of government 623

5.3.2. Differentiated subsidy strategy of government 669

Figure 10. 𝑆𝑊 𝐷𝐶∗ 𝑣𝑠. 𝑆𝑊 𝐷𝑀𝑆∗ 𝑣𝑠. 𝑆𝑊 𝐷𝑅𝑆∗ 720

Institutional Review Board Statement: Not applicable. 789

Informed Consent Statement: Not applicable. 790

Data Availability Statement: Not applicable. 791

Conflicts of Interest: The authors declare no conflict of interest. 792

Proof of Lemma 1. 794

Proof of Corollary 1: 808

𝜕𝑞2𝐶∗ 2𝑘(𝑠+𝜃−1−𝑐)𝜆 𝜕𝑞2𝐶∗ 𝑘 𝜕𝑒 𝐶∗ 2(𝑠+𝜃−1−𝑐)(𝜃−1)𝜆 𝜕𝑒 𝐶∗

Proof of Theorem 1: 819

Taking the second order derivative of 𝑆𝑊 𝐶 (𝑠) w.r.t. 𝑠, we have 825

Proof of Corollary 2: 832

Proof of Theorem 2: 834

Proof of Lemma 2: 837

Secondly, we solve the manufacturer’s optimization problem. We verify whether 845

Proof of Proposition 1: 861

Proof of Proposition 2: 867

You might also like