Product Innovation for Remanufacturing in the Closed-Loop Supply Chain with Third-Party Remanufacturing

Abstract

Product innovation for remanufacturing, beginning at the development stage, has become an important strategic decision in third-party remanufacturing. This study investigates decision-making on product innovation for remanufacturing under two third-party remanufacturing modes and examines how original equipment manufacturers (OEMs) and remanufacturers respond. Results show that outsourcing remanufacturing consistently offers a wider profitability range for the OEM and increases the likelihood of the remanufacturer adopting a full remanufacturing strategy. Furthermore, a higher innovation level enhances OEM profits, particularly when the remanufacturing industry is mature or when the innovation investment efficiency is high. Otherwise, incremental innovation is more beneficial. Innovation also lowers entry barriers for remanufacturers. Finally, the authorization remanufacturing is initially more environmentally friendly, whereas the outsourcing mode becomes superior as the innovation level increases.

1. Introduction

With the acceleration of economic development, industrial production activities have increasingly imposed significant environmental impacts. As a result, the sustainable circular economy is advocated as a replacement for the traditional production paradigm, which is considered unsustainable due to its high consumption and emissions [1]. Remanufactured products are shown to conserve raw materials and reduce pollution more effectively than new products [2]. Remanufacturing is widely regarded as an effective approach to promoting environmental preservation and resource conservation [3]. In China, efforts are underway to standardize the growth of the remanufacturing industry, alongside implementing special measures to encourage the circular economy. The remanufacturing process involves specific procedures that allow used products to achieve, or even surpass, the performance of new products [4]. Due to their high cost-effectiveness, remanufactured products are particularly popular in some developing countries. Owing to these benefits, remanufacturing has proliferated across various sectors, including the automotive and electronics industries [5,6].

Because of the lack of experience in remanufacturing, limited capital, corporate strategy, and other reasons, remanufacturing may not be a feasible choice for some original equipment manufacturers (OEMs) [7,8]. For instance, Ford once acquired several firms specializing in car and component recycling to launch a business in used vehicle disposal. However, the lack of sufficient expertise in this sector led Ford to eventually abandon this initiative and refocus on automaking [9]. Third-party entities may carry out remanufacturing activities for such companies. Given the substantial market demand for reused products, third-party remanufacturing has evolved into a profitable business [10]. Supported by government incentives, third-party remanufacturing has become a major mode in practice [11]. Third-party remanufacturing is widely adopted in sectors with high asset value and complex product structures. For instance, in the automotive industry, third-party firms are often responsible for the remanufacturing of critical components, such as engines, transmissions, and alternators. Similarly, in the consumer electronics sector, third-party entities play a pivotal role in the refurbishment of smartphones and printed circuit boards. OEMs may collaborate with remanufacturers through either authorization remanufacturing (AR) mode or outsourcing remanufacturing (OR) mode [12]. The major difference between AR mode and OR mode is who holds the rights to sell the remanufactured products. In the AR mode, the sale right is granted to the remanufacturer, as exemplified by HP and Lenovo, and the OEM sets the authorization fee for the remanufacturer [13]. The remanufactured products are marketed under the OEM’s brand name. Conversely, the OEM keeps the sale right and compensates the remanufacturer with an outsourcing fee for the production services in OR mode [14,15]. Caterpillar, for instance, offers Land Rover remanufacturing services. In third-party remanufacturing, there is also an independent remanufacturing (IR) mode, where the remanufacturer operates independently of the OEM and engages in remanufacturing activities [16]. IR mode typically emerges in markets where intellectual property protection is weak or regulatory oversight is lacking, and it may expose firms to a higher likelihood of infringement. In contrast, in environments with a well-established intellectual property protection system, AR and OR modes are more prevalent. Furthermore, the practical viability of these modes is increasingly supported by evolving business ecosystems and advanced technologies. For instance, additive manufacturing (3D printing) has emerged as a critical enabler for remanufacturing operations by reducing reverse supply chain complexity [17]. Moreover, digital technologies, such as blockchain and smart contracts, facilitate secure licensing by ensuring intellectual property protection and automating royalty payments [18]. These technological advancements significantly reduce the agency risks and transaction costs associated with authorization, thereby reinforcing the feasibility of AR and OR modes in the digital era. Therefore, considering intellectual property protection and regulatory compliance, this study focuses on AR and OR modes.

Some companies are integrating product innovation into new product development to enhance remanufacturing efficiency for environmental sustainability, resource conservation, and corporate image enhancement [19]. In recent years, the implementation of carbon neutrality goals and the Extended Producer Responsibility (EPR) system have driven a transformation in manufacturing. Under strict environmental regulations, OEMs must assume responsibility for the entire lifecycle of their products, particularly the post-consumer stage. Relying only on recovery and remanufacturing is no longer sufficient. Therefore, OEMs are required to consider remanufacturing strategies as early as the product development stage. Utilizing recyclable, eco-friendly materials and designing products for easy disassembly can significantly enhance remanufacturing capacity [20]. Considering remanufacturing from the source also aligns with the principles of the EPR system for enterprises [21]. This positive approach allows firms to not only meet emission reduction targets but also mitigate potential penalties associated with strict regulatory standards. This concept is referred to as product innovation for remanufacturing in this study. For instance, Coca-Cola replaced its iconic green Sprite bottles with transparent ones to facilitate reuse. In copier remanufacturing, Fuji Xerox implemented improvements that resulted in a three-fold reduction in waste costs. Nevertheless, a significant initial investment is necessary [22]. OEMs often perform a vital role in product innovation for remanufacturing. To improve remanufacturing efficiency, OEMs are required to make substantial investments in product development [23].

Although product innovation for remanufacturing benefits the remanufacturing industry, it has also brought changes to third-party remanufacturing. From a competitive production standpoint, product innovation for remanufacturing contributes to cost reductions, potentially enhancing the competitiveness of remanufactured products. In AR mode, competition between OEMs and remanufacturers intensifies significantly. In OR mode, OEMs sell both new and remanufactured products, creating a challenge in balancing product competition. Viewed from profits, engaging in product innovation for remanufacturing incurs additional costs for OEMs. Profitability stands as a key factor for OEMs. The ability to offset the costs of product innovation for remanufacturing with the resultant benefits significantly influences the decision-making of OEMs. Despite increased consumer willingness to purchase due to OEMs’ environmental initiatives [24], OEMs are not always motivated enough to innovate [20]. For remanufacturers, remanufacturing costs impact their strategies to enter the market [25], while product innovation for remanufacturing helps reduce the cost. The conventional view holds that high levels of product innovation in the forward supply chain can negatively impact the environment [26]. In the reverse supply chain, although product innovation for remanufacturing reduces the environmental impact of individual remanufactured products, the environmental impact also depends on the product quantity. Consequently, the extent to which innovation can reduce the environmental impact remains uncertain.

The role of product innovation in the forward supply chain has been explored. However, how product innovation for remanufacturing influences the decision-making of members within reverse supply chains has not been well studied. Furthermore, since reducing environmental impact is one of the purposes of remanufacturing, it is important to assess whether product innovation for remanufacturing can contribute to this objective. The relationship between OEMs and remanufacturers varies across different third-party remanufacturing modes, which in turn influences their decisions [27]. Thus, it is essential to examine the role of product innovation for remanufacturing in different modes. Only by addressing these issues can OEMs and remanufacturers make decisions effectively and thereby advance the development of remanufacturing.

Accordingly, this study focuses on the following research questions: (1) What are the attitudes of the OEM and remanufacturers toward product innovation for remanufacturing under third-party remanufacturing modes? How is the optimal level of product innovation for remanufacturing determined? (2) How does product innovation for remanufacturing influence the market entry strategies of the remanufacturer? How should the OEM and remanufacturer adjust their production and pricing decisions based on the innovation level? (3) Can product innovation for remanufacturing reduce environmental impact? What are the differences in its role under different remanufacturing modes?

To answer these research questions, a closed-loop supply chain involving an OEM and a remanufacturer is constructed. The OEM delegates the production of remanufactured products to the remanufacturer using AR or OR modes. The product innovation for remanufacturing of the OEM is comprehensively depicted under two modes. By studying the impact of the adoption of product innovation for remanufacturing on third-party remanufacturing, the optimal level of innovation is identified. The role of product innovation for remanufacturing in promoting remanufacturers to enter the market is also discussed. Additionally, the conditions under which product innovation for remanufacturing can reduce environmental impact are analyzed.

The remaining sections are structured as follows. Section 2 reviews the relevant literature. The detailed description and model assumptions of this study are provided in Section 3. Section 4 details the construction of two models. A detailed analysis is provided based on the results in Section 5. Finally, the conclusions and managerial implications are presented in Section 6.

2. Literature Review

Papers addressing remanufacturing modes and technological innovation are relevant to this research. A comprehensive review of the related literature is included.

2.1. Remanufacturing Mode

Extensive literature has explored the remanufacturing mode, providing valuable insights. While remanufacturing offers environmental and economic benefits, it may potentially pose challenges to OEMs due to the cannibalization effect [28]. This effect implies that remanufactured products may encroach upon the market. Some companies upgrade new products to differentiate them from remanufactured products [26,29]. However, many OEMs are not motivated to remanufacture because of the limitations of technology or the market [8]. Furthermore, third-party remanufacturing can lead to a series of patent disputes, raising concerns for OEMs [13]. To address these challenges, some OEMs choose to collaborate with remanufacturers by imposing an authorization charge [30]. This authorization cooperation between OEMs and remanufacturers becomes crucial when determining the authorization fee [27]. It enables OEMs to set suitable authorization fees to optimize profits and maintain market dominance in various market environments [31].

Additionally, some OEMs choose to outsource remanufacturing. While reducing the cost of collecting used products can mitigate the cannibalization effect, it may also lead to reduced profits for OEMs in the OR mode [32]. The OR mode is generally beneficial for OEMs [33]. Jin further discovers which mode is more advantageous for OEMs, depending on whether wholesale prices are endogenous [34]. When consumers perceive remanufactured products as high value, remanufacturers prefer the AR mode [35]. Zhou studies win–win strategies for OEMs and remanufacturers under different remanufacturing strategies [25]. In fact, independent remanufacturers handle a great deal of remanufacturing activity [36]. OEMs find the low pricing strategy more profitable, whereas independent remanufacturers benefit from a higher pricing strategy [37]. Regardless of whether independent remanufacturers are conducting remanufacturing, OEMs must maintain product quality. Engaging authorized remanufacturers is a strategy for OMEs to compete with unauthorized ones [27]. There may be both authorized and unauthorized remanufactured products in the market. Consumers often prefer authorized remanufacturers due to the guaranteed quality of their products [38].

2.2. Technology Innovation in the Closed-Loop Supply Chain

The consumer awareness regarding environmental sustainability incentivizes OEMs to innovate to produce greener products [39]. Technological innovation to facilitate remanufacturing often includes advancements in product and process innovation.

Product innovation for remanufacturing mainly involves improvements in product materials and design. The influence of product design on the environment and economy has been evaluated [40]. OEMs’ decisions are significantly influenced by cost savings, which are tightly linked to product design [41]. By adapting product design, OEMs can mitigate profit losses from remanufacturing and maintain a competitive advantage [42,43]. Controlling the remanufacturing cost also helps achieve win–win outcomes in the supply chain [44]. However, OEMs face challenges in determining the optimal level of product design disassemblability. In third-party remanufacturing, increased disassemblability, while reducing costs for remanufacturers, may adversely impact OEM profits [16,45]. Process innovation has been investigated to lower remanufacturing costs and enhance remanufacturing performance [46]. In addition, updating equipment can further facilitate development [47,48].

Generally, some OEMs have adopted technological innovation in remanufacturing to update products and expand market influence. Differing from process innovation, product innovation for remanufacturing not only lowers the costs of remanufacturing but also increases environmental sustainability. By using recyclable materials and easy-to-disassemble designs, product innovation can increase consumer purchase intent and align with the requirements of the EPR system. Since product innovation for remanufacturing needs to be considered during the product development stage, it is primarily carried out by OEMs. In addition to OEMs, companies involved in product remanufacturing can also undertake process innovation for remanufacturing. The extent to which product innovation for remanufacturing is beneficial directly influences the likelihood of OEMs pursuing innovation. As product innovation can reduce the cost of remanufactured products, the degree of innovation by OEMs plays a crucial role in shaping the market entry strategies of remanufacturers. This, in turn, has a significant impact on subsequent production decisions.

In practice, third-party remanufacturing has become a common remanufacturing mode. Stepwise process innovation is required in remanufacturing and has the potential to reduce environmental impact [19]. Other members can share the cost of innovation to facilitate OEMs’ process innovation [23]. Despite these advancements, existing research on product and process innovation for remanufacturing has primarily focused on scenarios where the OEM conducts remanufacturing internally, as shown in

Table 1. Factors considered in the related literature and this paper.

Papers	Third-Party Remanufacturing			OEM Remanufacturing	Product Innovation for Remanufacturing	Process Innovation for Remanufacturing
	AR	OR	IR
Reimann et al. [19]				√		√
Chen et al. [23]				√		√
Chai et al. [46]				√		√
Wu [16]			√		√
Qiang et al. [45]				√	√
Liu et al. [42]				√	√
Gu [24]				√	√
Wang et al. [20]		√			√
Zou et al. [35]	√	√
This paper	√	√			√

. Zou et al. [35] compared the AR and OR modes without considering product innovation. However, with the development of the remanufacturing industry, product innovation for remanufacturing has become a critical requirement driven by policy orientation and environmental sustainability demands. Furthermore, it also contributes to a superior corporate image. There remains a notable absence of theoretical analysis that integrates product innovation for remanufacturing with third-party remanufacturing, specifically within the contexts of AR and OR modes. Furthermore, the specific environmental implications of product innovation for remanufacturing within these modes are often overlooked. Therefore, this study incorporates product innovation for remanufacturing into AR and OR scenarios to explore how innovation levels influence the decision-making and profitability of supply chain members. This study also assesses whether such innovation genuinely contributes to environmental sustainability, thereby offering robust recommendations for the development of the remanufacturing industry.

3. Problem Description and Assumptions

A closed-loop supply chain comprising an OEM and a remanufacturer is considered, where both new and remanufactured products are available in the market. The OEM is responsible for the development and marketing of new products and engages in product innovation for remanufacturing during the design step. For example, using an easy-to-disassemble design, and using recyclable or environmentally friendly materials. Remanufacturing is conducted through third-party remanufacturing modes. Assume that the decision process occurs within a single period to exclude the influence of the start and terminal periods. The decisions remain stable during this period [28,46].

The relevant variable symbols are summarized in

Table 2. The notations in this study.

Parameters
$c_n/c_r$	Unit cost of new/remanufactured products
$\beta$	Recovery cost coefficient
$\lambda$	The level of product innovation for remanufacturing
k	Innovation cost coefficient
$v$	The willingness of consumers to buy a new product
$u_n/u_r$	The net utility of purchasing new/remanufactured products
$\theta$	Consumer’s value discount coefficient for remanufactured products, $0<\theta <1$
$\alpha$	Environmental impact of the remanufactured product (new product = 1)
${\pi }_n/{\pi }_r$	Profit of the OEM/remanufacturer
*	A superscript denoting the optimal value
L	A superscript denoting a different model, L = (AR, OR)
Decision variables
$q_n/q_r$	Quantity of new/remanufactured products
$p_n/p_r$	Unit price of new/remanufactured products
$\gamma$	Recovery rate
$\epsilon$	Unit authorization fee
$\rho$	Unit outsourcing fee

.

3.1. The Structure of Third-Party Remanufacturing

There are two modes for third-party remanufacturing: AR mode (Figure 1a) and OR mode (Figure 1b). In the Stackelberg game, the OEM and the remanufacturer serve as the leader and the follower, respectively.

Under the AR mode, the OEM authorizes the remanufacturer to produce and market remanufactured products, with the remanufacturer paying an authorization fee. The OEM first decides quantity of new products $q_n$ and unit authorization fee $\epsilon$. Then, the remanufacturer determines recovery rate $\gamma$ and quantity of remanufactured products $q_r$. In the OR mode, the OEM pays an outsourcing charge to the remanufacturer to develop the remanufactured product. The remanufactured products will be delivered to the OEM after production is finished. Following the OEM’s determination of $q_n$ and unit outsourcing fee $\rho$, the remanufacturer decides $\gamma$ and $q_r$. The decision-making sequence for both modes is shown in Figure 2.

3.2. The Cost Structure

Remanufactured products can conserve energy and resources compared to new products, so the unit cost of remanufactured products is lower ($c_n>c_r$) [19]. The product innovation for remanufacturing contributes to enhanced remanufacturability of new products, making them easier to disassemble and increasing the potential for material reuse, which in turn reduces remanufacturing costs. The extent of cost reduction depends on the OEM’s level of product innovation for remanufacturing ($\lambda$). If $\lambda =0$, the OEM does not carry out product innovation for remanufacturing. Indeed, the unit cost of remanufactured products is established as $c_r-\lambda$ [23]. To depict the diminishing returns from innovation investment, the cost of product innovation for remanufacturing is $k{\lambda }^2/2$ [49]. Moreover, k can also reflect the innovation capability of the OEM to some extent [50].

In reality, only a portion of end-of-life products can be recovered and remanufactured. However, consistent with assumptions commonly made in the literature [19], it is assumed that all end-of-life products are collectible, and all recycled products are eligible for remanufacturing [19,23]. The end-of-life product can be remanufactured once [51]. The recovery cost is modeled as a convex function. The recovery cost is represented as $\beta {(\gamma q_n)}^2/2$, where $\beta$ reflects the recovery cost coefficient.

3.3. Inverse Demand Functions

The market capacity is standardized to 1. Assume that v follows a uniform distribution ranging from 0 to 1. After the product innovation for remanufacturing, the green degree of new products is enhanced, thereby increasing consumer purchase willingness to $(1+\lambda )\nu$ [24]. Although remanufactured products and new products are of equivalent quality [52], consumers have different purchasing preferences. The consumer’s value discount coefficient for remanufactured products is denoted as $\theta$. Therefore, the purchasing intention of consumers for remanufactured products is $(1+\lambda )\nu \theta$ [38].

Consumers’ net utility is influenced by both price ($p_n$, $p_r$) and innovation level. In AR and OR modes, the net utility of consumers buying two types of products, respectively, is given by $u_n=(1+\lambda )\nu -p_n$ and $u_r=(1+\lambda )\nu \theta -p_r$ [46]. Each consumer buys no more than one product. There are three choices for consumers. Consumers purchase the product with the highest net utility if it is greater than zero. When $u_n<0$ and $u_r<0$, the consumers will give up buying. There is no information asymmetry. It is assumed that all production schedules can be determined based on the market demand, resulting in no inventory [19]. The inverse demand functions for two types of products, respectively, are delineated as $p_n=(1+\lambda )(1-q_n-\theta q_r)$ and $p_r=(1+\lambda )\theta (1-q_n-q_r)$. Meanwhile, we can get $q_r=\gamma q_n$. Unit prices for two types of products can be represented as functions of the quantity of each product [23]. The amount of remanufactured products can be further transformed into a function of the recovery rate [35].

4. Mathematical Models

The models in two remanufacturing modes are constructed in this section. Both the OEM and the remanufacturer pursue profit maximization. The optimal values are derived using the method of backward induction, with detailed calculations presented in Appendix A.

4.1. AR Mode

The OEM grants the remanufacturer permission to produce and sell remanufactured products in exchange for an authorization fee [35]. After the product innovation for remanufacturing, the OEM decides on $q_n$ and $\epsilon$ firstly. Then, according to the decision of the OEM, the remanufacturer sets $\gamma$ for gathering discarded products.

The OEM’s profit function is defined as

$$\underset{q_n,\epsilon }{\max } {\pi }_n^{AR}=(p_n-c_n)q_n+\epsilon q_r-k{\lambda }^2/2,$$

(1)

where the first part consists of revenues from new products, the next part is the authorization fee, and the final part covers innovation costs.

The remanufacturer’s profit function is determined by

$$\underset{\gamma }{\max } {\pi }_r^{AR}=(p_r-c_r+\lambda )q_r-\epsilon q_r-\beta {(\gamma q_n)}^2/2,$$

(2)

where the initial part includes the profit generated by remanufactured products, the next part is the authorization fee, and the third part is recovery costs.

Proposition 1 summarizes the market entry strategy of the remanufacturer in the AR mode.

Proposition 1.

There exist two thresholds of ${\overline{c}}_r^{AR}=\min \{c_n(1+\lambda )\theta +\lambda ,c_n\}$ and  ${\underset{\_}{c}}_r^{AR}=\max \{{\displaystyle \frac{(1+\lambda ){\theta }^2-((1-3c_n)\lambda -3c_n+2)\theta +\lambda -(1-c_n)\beta }{\theta +1}},\lambda \}$  such that the remanufacturing strategy of the remanufacturer is:

(1) 

No remanufacturing when $c_n>c_r>{\overline{c}}_r^{AR}$,

(2) 

Partial remanufacturing when ${\overline{c}}_r^{AR}>c_r>{\underset{\_}{c}}_r^{AR}$,

(3) 

Full remanufacturing when ${\underset{\_}{c}}_r^{AR}>c_r>\lambda$.

The decision for remanufacturers to enter the market is largely influenced by the remanufacturing cost. When $c_r>{\overline{c}}_r^{AR}$, the cost of remanufacturing is so high that the remanufacturers may not engage in remanufacturing ($q_r^{AR*}=0$). When ${\underset{\_}{c}}_r^{AR}<c_r<{\overline{c}}_r^{AR}$, remanufacturing becomes profitable as the costs decrease. The remanufacturers start collecting discarded products for remanufacturing ($q_r^{AR*}<q_n^{AR*}$). When $c_r<{\underset{\_}{c}}_r^{AR}$, the remanufacturers are encouraged to collect all discarded products for remanufacturing ($q_r^{AR*}=q_n^{AR*}$).

Depending on the possible values of ${\overline{c}}_r^{AR}$ and ${\underset{\_}{c}}_r^{AR}$, the remanufacturing strategy space of the remanufacturer can be divided into different cases, as shown in

Table 3. Remanufacturing strategy space under different conditions in the AR mode.

Conditions			Remanufacturing Strategy
$\theta +3c_n-2<0$		$\lambda <{\lambda }_1^{AR}$	No, Partial
		$\lambda >{\lambda }_1^{AR}$	Partial
$\theta +3c_n-2>0$	$\beta <{\beta }_1^{AR}$	$\lambda <{\lambda }_1^{AR}$	No, Partial, Full
		${\lambda }_1^{AR}<\lambda <{\lambda }_2^{AR}$	Partial, Full
		$\lambda >{\lambda }_2^{AR}$	Full
	${\beta }_1^{AR}<\beta <{\beta }_2^{AR}$	$\lambda <{\lambda }_1^{AR}$	No, Partial
		${\lambda }_1^{AR}<\lambda <{\lambda }_3^{AR}$	Partial
		${\lambda }_3^{AR}<\lambda <{\lambda }_2^{AR}$	Partial, Full
		$\lambda >{\lambda }_2^{AR}$	Full
	$\beta >{\beta }_2^{AR}$	$\lambda <{\lambda }_1^{AR}$	No, Partial
		$\lambda >{\lambda }_1^{AR}$	Partial

, where ${\lambda }_1^{AR}={\displaystyle \frac{c_n(1-\theta )}{\theta c_n+1}}$, ${\lambda }_2^{AR}={\displaystyle \frac{-{\theta }^2+2(1-c_n)\theta +c_n+(1-c_n)\beta }{{\theta }^2+(3c_n-1)\theta +1}}$, ${\lambda }_3^{AR}={\displaystyle \frac{(1-c_n)\beta -(\theta +3c_n-2)\theta }{\theta (\theta +3c_n-2)}}$, ${\beta }_1^{AR}={\displaystyle \frac{\theta (\theta +3c_n-2)}{1-c_n}}$, and ${\beta }_2^{AR}={\displaystyle \frac{\theta (1+c_n)(\theta +3c_n-2)}{1-c_n}}$.

Remanufacturing cost plays a decisive role in the remanufacturer’s market entry decision, which is influenced by the product innovation level. When $\lambda >{\lambda }_1^{AR}$, which means that the remanufacturing cost has decreased sufficiently, the remanufacturer will carry out remanufacturing. Whether the remanufacturer has the option to adopt a full remanufacturing strategy mainly depends on consumer acceptance of remanufactured products, the recovery cost coefficient, and the innovation level. A low level of consumer acceptance of remanufactured products makes the adoption of a full remanufacturing strategy by the remanufacturer unlikely. Compared to new products, remanufactured products have lower competitiveness. As a result, the remanufacturer may not enter the market at all or may only collect a portion of discarded products to remanufacture.

Greater consumer willingness to purchase remanufactured products leads to higher market acceptance of remanufacturing. However, the recovery cost coefficient directly impacts the recovery cost. Rising recovery costs can be mitigated by higher innovation levels, which help maintain the remanufacturer’s incentive to adopt a full remanufacturing strategy. Even after the innovation level has been enhanced to a certain level, the remanufacturer will adopt a full remanufacturing strategy. Nevertheless, once the recovery cost reaches a certain threshold, even if the innovation level is enhanced to lower the remanufacturing cost, the remanufacturer is unlikely to adopt a full remanufacturing strategy. The remanufacturing strategy space of the remanufacturer is shown in Figure 3, based on four sets of consumer preferences and recycling cost coefficients.

As the innovation level increases, remanufacturers can accept higher initial remanufacturing costs to enter the market and are more likely to adopt full remanufacturing ($\partial {\underset{\_}{c}}_r^{AR}/\partial \lambda >0$). There exist only new products, with no remanufactured items available under the no remanufacturing strategy. Therefore, this study only focuses on the boundaries of the no remanufacturing strategy, without addressing the decisions of the OEM under this strategy.

Proposition 2.

The optimal decisions of the OEM and remanufacturer in the AR mode are given in

Table 4. Optimal decisions in the AR mode.

Region	${\underset{\mathbf{\_}}{\mathit{c}}}_{\mathit{r}}^{\mathit{A}\mathit{R}}\mathbf{<}{\mathit{c}}_{\mathit{r}}\mathbf{<}{\overline{\mathit{c}}}_{\mathit{r}}^{\mathit{A}\mathit{R}}$	${\mathit{c}}_{\mathit{r}}\mathbf{<}{\underset{\mathbf{\_}}{\mathit{c}}}_{\mathit{r}}^{\mathit{A}\mathit{R}}$
$q_n^{AR*}$	${\displaystyle \frac{(\lambda +c_r+2)\theta -(\lambda +1)(\theta +2c_n)\theta -\beta (c_n-1)}{-2(\lambda +1){\theta }^2+4(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{(\lambda +c_r+2)\theta -(\lambda +1)(\theta +c_n)\theta +\lambda -c_r-\beta (c_n-1)}{-4(\lambda +1){\theta }^2+8(\lambda +1)\theta +4\beta }}$
${\epsilon }^{AR*}$	${\displaystyle \frac{(\lambda +1)\theta +\lambda -c_r}{2}}$	${\displaystyle \frac{\begin{array}{l}3{\theta }^2{(\lambda +1)}^2(3c_n-\theta )-3(5\theta (\lambda +1)+3\beta )(c_r-\lambda )+3(c_n\\ -1){\beta }^2+3\theta \beta (4c_n(\lambda +1)-c_r-1)+{\theta }^2(c_r+\lambda +2+\beta )\end{array}}{12(-(\lambda +1){\theta }^2+2(\lambda +1)\theta +\beta }}$
$q_r^{AR*}$	${\displaystyle \frac{c_n(\lambda +1)\theta +\lambda -c_r}{-2(\lambda +1){\theta }^2+4(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{(\lambda +c_r+2)\theta -(\lambda +1)(\theta +c_n)\theta +\lambda -c_r-\beta (c_n-1)}{-4(\lambda +1){\theta }^2+8(\lambda +1)\theta +4\beta }}$
${\gamma }^{AR*}$	${\displaystyle \frac{c_n(\lambda +1)\theta +\lambda -c_r}{(\lambda +c_r+2)\theta -(\lambda +1)(\theta +2c_n)\theta -\beta (c_n-1)}}$	1
$p_n^{AR*}$	${\displaystyle \frac{(\lambda +1)(c_n+1)}{2}}$	${\displaystyle \frac{\begin{array}{l}(\lambda +1)(\theta (\lambda +1)(\theta -2)(\theta -3)+c_n(\theta \lambda +\beta +\theta )\\ (\theta +1)+(c_r-\lambda )(1-{\theta }^2)+\beta (3-\theta ))\end{array}}{-4(\lambda +1){\theta }^2+8(\lambda +1)\theta +4\beta }}$
$p_r^{AR*}$	${\displaystyle \frac{\theta (\lambda +1)((\lambda +1)(\beta c_n+(c_n+2)\theta -{\theta }^2)+(c_r-\beta )(1-\theta ))}{-2(\lambda +1){\theta }^2+4(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{\theta (\lambda +1)((\lambda +1)(\beta c_n+(c_n+2)\theta -{\theta }^2)+(c_r-\beta )(1-\theta ))}{-2(\lambda +1){\theta }^2+4(\lambda +1)\theta +2\beta }}$

.

4.2. OR Mode

The OEM pays an outsourcing charge to the remanufacturer to produce remanufactured products [35]. After the product innovation for remanufacturing, the OEM sets $q_n$ and $\rho$ firstly. Given the decision of the OEM, the remanufacturer decides $\gamma$ subsequently.

The OEM’s profit function is defined as

$$\underset{q_n,\rho }{\max } {\pi }_n^{OR}=(p_n-c_n)q_n+(p_r-\rho )q_r-k{\lambda }^2/2,$$

(3)

where the first part includes the profit generated by new products, the next part comprises the earnings from selling remanufactured products minus outsourcing fees, and the third part is innovation costs.

The remanufacturer’s profit function is determined by

$$\underset{\gamma }{\max } {\pi }_r^{OR}=(\lambda +\rho -c_r)q_r-\beta {(\gamma q_n)}^2/2,$$

(4)

where the first part is the income from remanufactured products and the outsourcing fee. The second part is the recovery cost.

Proposition 3 summarizes the remanufacturing strategy of the remanufacturer in the OR mode.

Proposition 3.

There exist two thresholds of ${\overline{c}}_r^{OR}=\min \{c_n(1+\lambda )\theta +\lambda ,c_n\}$ and  ${\underset{\_}{c}}_r^{OR}=\max \{{\displaystyle \frac{(1+\lambda ){\theta }^2+(2c_n(\lambda +1)-1)\theta +\lambda -(1-c_n)\beta }{\theta +1}},\lambda \}$  such that the remanufacturing strategy of the remanufacturer is:

(1) 

No remanufacturing when $c_n>c_r>{\overline{c}}_r^{OR}$,

(2) 

Partial remanufacturing when ${\overline{c}}_r^{OR}>c_r>{\underset{\_}{c}}_r^{OR}$,

(3) 

Full remanufacturing when ${\underset{\_}{c}}_r^{OR}>c_r>\lambda$.

Similar results to those under the AR mode can be obtained. The strategy of the remanufacturer to engage in remanufacturing is also primarily influenced by remanufacturing costs. Depending on the possible values of ${\overline{c}}_r^{OR}$ and ${\underset{\_}{c}}_r^{OR}$, the remanufacturing strategy space of the remanufacturer can be divided into different cases, as shown in

Table 5. Remanufacturing strategies corresponding to different conditions in the OR mode.

Conditions			Remanufacturing Strategy
$\theta +2c_n-1<0$		$\lambda <{\lambda }_1^{OR}$	No, Partial
		$\lambda >{\lambda }_1^{OR}$	Partial
$\theta +2c_n-1>0$	$\beta <{\beta }_1^{OR}$	$\lambda <{\lambda }_1^{OR}$	No, Partial, Full
		${\lambda }_1^{OR}<\lambda <{\lambda }_2^{OR}$	Partial, Full
		$\lambda >{\lambda }_2^{OR}$	Full
	${\beta }_2^{OR}>\beta >{\beta }_1^{OR}$	$\lambda <{\lambda }_1^{OR}$	No, Partial
		${\lambda }_1^{OR}<\lambda <{\lambda }_3^{OR}$	Partial
		${\lambda }_3^{OR}<\lambda <{\lambda }_2^{OR}$	Partial, Full
		$\lambda >{\lambda }_2^{OR}$	Full
	$\beta >{\beta }_2^{OR}$	$\lambda <{\lambda }_1^{OR}$	No, Partial
		$\lambda >{\lambda }_1^{OR}$	Partial

, where ${\lambda }_1^{OR}={\displaystyle \frac{c_n(1-\theta )}{\theta c_n+1}}$, ${\lambda }_2^{OR}={\displaystyle \frac{-{\theta }^2+(1-c_n)\theta +c_n+(1-c_n)\beta }{{\theta }^2+2c_n\theta +1}}$, ${\lambda }_2^{OR}={\displaystyle \frac{-{\theta }^2+(1-c_n)\theta +c_n+(1-c_n)\beta }{{\theta }^2+2c_n\theta +1}}$, ${\beta }_1^{OR}={\displaystyle \frac{\theta (\theta +2c_n-1)}{1-c_n}}$, and ${\beta }_2^{OR}={\displaystyle \frac{\theta (1+c_n)(\theta +2c_n-1)}{1-c_n}}$.

In the OR mode, the strategy space of the remanufacturer is shown in Figure 4, based on four sets of consumer preferences and recycling cost coefficients.

Corollary 1.

The relationship between ${\overline{c}}_r^{AR}$ and  ${\overline{c}}_r^{OR}$, ${\underset{\_}{c}}_r^{AR}$ and  ${\underset{\_}{c}}_r^{OR}$ is as follows:  ${\overline{c}}_r^{AR}={\overline{c}}_r^{OR}$, ${\underset{\_}{c}}_r^{AR}\le {\underset{\_}{c}}_r^{OR}$.

Both modes impose the same entry conditions on the remanufacturer. However, the threshold for adopting full remanufacturing is lower under the AR mode. The remanufacturer must manage the sales of remanufactured products and face competition from the new products. Only with a lower remanufacturing cost will the remanufacturer be incentivized to pursue full remanufacturing.

It is easy to prove that $\theta +3c_n-2<\theta +2c_n-1$, ${\lambda }_1^{AR}={\lambda }_1^{OR}$, ${\lambda }_3^{AR}>{\lambda }_3^{OR}$, ${\lambda }_2^{AR}>{\lambda }_2^{OR}$, ${\beta }_1^{AR}<{\beta }_1^{OR}$, and ${\beta }_2^{AR}<{\beta }_2^{OR}$. Comparing the coefficient ranges including the full remanufacturing strategy, it is found that the range is larger in the OR mode. Therefore, the OR mode increases the likelihood that the remanufacturer will pursue a full remanufacturing strategy.

Proposition 4.

The optimal decisions of the OEM and remanufacturer in the OR mode are given in

Table 6. Optimal decisions in the OR mode.

Region	${\underset{\mathbf{\_}}{\mathit{c}}}_{\mathit{r}}^{\mathit{O}\mathit{R}}\mathbf{<}{\mathit{c}}_{\mathit{r}}\mathbf{<}{\overline{\mathit{c}}}_{\mathit{r}}^{\mathit{O}\mathit{R}}$	${\mathit{c}}_{\mathit{r}}\mathbf{<}{\underset{\mathbf{\_}}{\mathit{c}}}_{\mathit{r}}^{\mathit{O}\mathit{R}}$
$q_n^{OR*}$	${\displaystyle \frac{(c_r+1)\theta -(c_n+\theta )(\lambda +1)\theta -\beta (c_n-1)}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{(1-\theta )(\theta (\lambda +1)+\lambda -c_r)-\beta (c_n-1)}{-4(\lambda +1){\theta }^2+4(\lambda +1)\theta +4\beta }}$
${\rho }^{OR*}$	${\displaystyle \frac{(2\theta (1-(\lambda +1)\theta )+\beta )(c_r-\lambda )+(\lambda +1)\theta \beta c_n}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{\begin{array}{l}4\theta (\lambda +1)(1-\theta )(c_r-\lambda )-\beta ((\lambda +1){\theta }^2\\ -(c_r+1)\theta +\beta (c_n-1)+3\lambda -3c_r\end{array}}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$
$q_r^{OR*}$	${\displaystyle \frac{c_n(\lambda +1)\theta +\lambda -c_r}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{(1-\theta )(\theta (\lambda +1)+\lambda -c_r)-\beta (c_n-1)}{-4(\lambda +1){\theta }^2+4(\lambda +1)\theta +4\beta }}$
${\gamma }^{OR*}$	${\displaystyle \frac{c_n(\lambda +1)\theta +\lambda -c_r}{(c_r+1)\theta -(c_n+\theta )(\lambda +1)\theta -\beta (c_n-1)}}$	1
$p_n^{OR*}$	${\displaystyle \frac{(\lambda +1)(c_n+1)}{2}}$	${\displaystyle \frac{\begin{array}{l}(\lambda +1)(\theta (\lambda +1)(\theta -1)(\theta -3)+\beta ((c_n\\ -1)\theta +c_n+3)+\theta (c_r-\lambda )(1-\theta ))\end{array}}{-4(\lambda +1){\theta }^2+4(\lambda +1)\theta +4\beta }}$
$p_r^{OR*}$	${\displaystyle \frac{\theta (\lambda +1)((1-\theta )(\theta \lambda +c_r-\lambda +\theta )+(c_n+1)\beta )}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$	${\displaystyle \frac{\theta (\lambda +1)((1-\theta )(\theta \lambda +c_r-\lambda +\theta )+(c_n+1)\beta )}{-2(\lambda +1){\theta }^2+2(\lambda +1)\theta +2\beta }}$

.

5. Results and Discussion

A series of analyses using the results are conducted to answer the research questions. The related proof is included in Appendix A. To validate the theoretical propositions, a numerical analysis is conducted. The parameter values are based on relevant literature [19,23,35,46], with the data in Chai et al. [46] derived from practices in the electric vehicle industry. To ensure the solvability of the model and the existence of equilibrium, the parameters are set as follows: $c_n=0.4$, $c_r=0.35$, $\theta =0.9$, and $\beta =0.1$.

5.1. The Optimal Level of Product Innovation for Remanufacturing

The influence of $\lambda$ on the profits is analyzed to discover the optimal level of product innovation for remanufacturing.

Proposition 5.

The impact of $\lambda$ on the profits of the OEM and remanufacturer is as follows:

(1) 

For the OEM, there exist ${\widehat{k}}^L$ and ${\widehat{\lambda }}^L$ such that when $k<{\widehat{k}}^L$, ${\displaystyle \frac{\partial {\pi }_n^{L*}}{\partial \lambda }}>0$. When $k>{\widehat{k}}^L$, if $\lambda <{\widehat{\lambda }}^L$, ${\displaystyle \frac{\partial {\pi }_n^{L*}}{\partial \lambda }}>0$; otherwise, ${\displaystyle \frac{\partial {\pi }_n^{L*}}{\partial \lambda }}<0$.

(2) 

For the remanufacturer, ${\displaystyle \frac{\partial {\pi }_r^{L*}}{\partial \lambda }}>0$ always holds.

The relationship between ${\pi }_n^{L*}$ and $\lambda$ is influenced by k, as shown in Figure 5 and Figure 6. When k is small, ${\pi }_n^{L*}$ increases with respect to $\lambda$. The expenses incurred from innovation are compensated by the profits gained from remanufacturing or increased demand. Ultimately, the profits of the OEM even show improvement. When k is higher, an initial increase in profits is followed by a decrease as $\lambda$ rises. A higher innovation level begins to negatively impact the OEM’s profits, as the benefits from innovation cannot cover the increased cost. Therefore, whether innovation will be profitable for the OEM is influenced by both k and $\lambda$. The profitability, in turn, affects the OEM’s motivation toward innovation. From the perspective of the enterprise, k reflects the investment efficiency, which is influenced by factors such as management level and R&D input [28]. A smaller k indicates higher investment efficiency, meaning that lower cost is needed to achieve a higher level of innovation. From the industry perspective, k reflects the development stage of the industry. At the early stages of industry development, significant investment is often required to achieve technology innovation. As the industry gradually matures, k progressively decreases [23]. At different stages of industry development or under varying levels of investment efficiency of the OEM, the OEM may not always engage in process innovation [19]. However, there is always an optimal level of product innovation that benefits the OEM. The optimal level of innovation needs to be determined by k (the industry’s development stage or investment efficiency). When $k<{\widehat{k}}^L$, the optimal level of innovation is sufficiently high. When $k>{\widehat{k}}^L$, the optimal level of innovation is ${\widehat{\lambda }}^L$, which decreases with the increase of k.

For the remanufacturer, the product innovation for remanufacturing is always beneficial, as shown in Figure 7. What is more, the higher the innovation level, the greater the profits. In addition, regardless of which strategy is adopted by the remanufacturer, the impact trend remains consistent in two modes. However, under the OR mode, the range of k that always increases the OEM’ profit is larger (${\widehat{k}}^{OR}>{\widehat{k}}^{AR}$). Moreover, ${\widehat{\lambda }}^{OR}>{\widehat{\lambda }}^{AR}$. This implies that the scenarios in which innovation can bring profits to OEMs are more extensive under the OR mode.

5.2. Impact of Product Innovation for Remanufacturing on the Optimal Decisions

The product price is derived from the inverse demand function. By differentiating the optimal decisions in relation to the innovation level, Proposition 6 is established.

Proposition 6.

The impact of $\lambda$ on the optimal decisions is as follows:

(1) 

When the remanufacturer adopts partial remanufacturing,

(i) 

for the OEM, ${\displaystyle \frac{\partial q_n^{L*}}{\partial \lambda }}<0$, ${\displaystyle \frac{\partial {\epsilon }^{AR*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial {\rho }^{OR*}}{\partial \lambda }}<0$, ${\displaystyle \frac{\partial p_n^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial p_r^{OR*}}{\partial \lambda }}>0$;

(ii) 

for the remanufacturer, ${\displaystyle \frac{\partial q_r^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial {\gamma }^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial p_r^{AR*}}{\partial \lambda }}>0$.

(2) 

When the remanufacturer adopts full remanufacturing,

(i) 

for the OEM, ${\displaystyle \frac{\partial q_n^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial {\epsilon }^{AR*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial {\rho }^{OR*}}{\partial \lambda }}<0$, ${\displaystyle \frac{\partial p_n^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial p_r^{OR*}}{\partial \lambda }}>0$;

(ii) 

for the remanufacturer, ${\displaystyle \frac{\partial q_r^{L*}}{\partial \lambda }}>0$, ${\displaystyle \frac{\partial {\gamma }^{L*}}{\partial \lambda }}=0$, ${\displaystyle \frac{\partial p_r^{AR*}}{\partial \lambda }}>0$.

A higher innovation level correlates with reduced costs in remanufacturing. Consequently, the OEM may lower the outsourcing fees to reduce expenses or increase the authorization fee to raise income. The remanufacturer indirectly assists the OEM in sharing the cost of innovation. When the remanufacturer adopts partial remanufacturing, $q_n^{L*}$ and $q_r^{L*}$ change in the opposite direction as $\lambda$ rises. The remanufacturer will choose to recycle more used products for remanufacturing. Even if the innovation stimulates the willingness of consumers to buy, $q_n^{L*}$ might decrease. However, as all used products become recyclable and remanufacturable, the expanding market for remanufactured products can also lead to higher output of new products, intensifying the competition between the two types of products.

Note that the pricing trends of the two types of products move in the same direction. Consumers who prefer new products consequently need to share more of the cost. Although the cost of remanufacturing goes down, $p_r^{L*}$ increases. The prices of new and remanufactured products will be raised simultaneously. However, new products consistently have a higher price than remanufactured products ($p_n^{OR*}>p_r^{OR*}$); otherwise, the remanufactured products will become unsellable.

5.3. Environmental Impacts

Following related literature in closed-loop supply chains [26,29], the environmental impact can be investigated through the total material consumption. While the environmental impact is multidimensional, material consumption is widely regarded as a primary indicator in analytical modeling. This is because the production is typically the most energy-intensive and polluting stage in a product’s lifecycle. Therefore, material consumption provides a reasonable and tractable measure for environmental performance. Compared with new products, remanufactured products use fewer materials. Assuming the environmental impact of new products is set at 1, the environmental impact of remanufactured products is denoted as $\alpha (0<\alpha <1)$. After the innovation, the environmental impact of remanufactured products is further reduced to $\alpha -\lambda$. Therefore, considering the reduction due to innovation, the environmental impact is $E^{L*}=q_n^{L*}+(\alpha -\lambda )q_r^{L*}$.

Proposition 7.

The environmental impacts in the AR and OR modes satisfy the following relationships:

(1) 

When ${\overline{c}}_r^{AR}({\overline{c}}_r^{OR})>c_r>{\underset{\_}{c}}_r^{OR}$, if $\alpha >\theta +\lambda$, $E^{OR*}>E^{AR*}$; otherwise, $E^{AR*}>E^{OR*}$.

(2) 

When ${\underset{\_}{c}}_r^{OR}>c_r>{\underset{\_}{c}}_r^{AR}$, there exists $\widehat{\alpha }$ such that when $\alpha >\widehat{\alpha }$, $E^{OR*}>E^{AR*}$; otherwise, $E^{AR*}>E^{OR*}$.

(3) 

When ${\underset{\_}{c}}_r^{AR}>c_r$, $E^{OR*}>E^{AR*}$.

Only when the environmental savings from fewer new products exceed the added impact of increased remanufacturing can total environmental impact be reduced. When $c_r>{\underset{\_}{c}}_r^{AR}$, the supply of new products in the AR mode is higher than that of OR mode ($q_n^{AR*}>q_n^{OR*}$), while the supply of remanufactured products is lower than that of OR mode ($q_r^{AR*}<q_r^{OR*}$). A comparison with Zou et al. [35] reveals that the comparative environmental impacts of the two modes depend not only on consumer acceptance of remanufactured products but are also significantly influenced by the innovation level and the degree of material savings. Furthermore, the environmental comparison yields distinct results across the TPR’s different market entry strategies.

When ${\overline{c}}_r>c_r>{\underset{\_}{c}}_r^{OR}$, the remanufacturer adopts partial strategy in both modes. If $\alpha >\theta +\lambda$, remanufactured products deliver a smaller reduction in environmental impact. Consequently, the OR mode results in a higher environmental impact. Conversely, if the reduction in the environmental impact of remanufactured products is more significant, the increased environmental impact of AR mode will become higher, as shown in Figure 8a. As $\lambda$ increases, the threshold of $\alpha$ becomes larger. There are more scenarios to make OR mode greener.

When ${\underset{\_}{c}}_r^{OR}>c_r>{\underset{\_}{c}}_r^{AR}$, the AR mode leads the remanufacturer to adopt a partial strategy, while the OR mode induces a full strategy. As $\alpha$ increases, first the OR mode is greener, then the AR mode, as shown in Figure 8b. However, the threshold of $\alpha$ decreases as $\lambda$ increases. In the OR mode, higher values of $\lambda$ result in increased production of both new and remanufactured products, whereas in the AR mode, production shifts from new products to remanufactured products. There are more scenarios to make AR mode greener.

When ${\underset{\_}{c}}_r^{AR}>c_r$, the remanufacturer adopts the full strategy in both modes. The OR mode results in larger outputs of both new and remanufactured products than the AR mode, thereby generating greater environmental impact, as depicted in Figure 8c.

Proposition 8.

The impact of $\lambda$ on environmental impacts is as follows:

(1) 

When the remanufacturer adopts partial remanufacturing, there exists ${\theta }_1^L$ such that when $\theta <{\theta }_1^L$, ${\displaystyle \frac{\partial E^{L*}}{\partial \lambda }}>0$; otherwise, ${\displaystyle \frac{\partial E^{L*}}{\partial \lambda }}<0$.

(2) 

When the remanufacturer adopts full remanufacturing, there exists ${\theta }_2^L$ such that when $\theta <{\theta }_2^L$, ${\displaystyle \frac{\partial E^{L*}}{\partial \lambda }}>0$; otherwise, ${\displaystyle \frac{\partial E^{L*}}{\partial \lambda }}<0$.

Innovation can only benefit the environment when the acceptance of consumers to remanufacturing is sufficiently high. This is often challenging to achieve in the early stage of remanufacturing development. Therefore, innovation in the early stage of remanufacturing development often comes at the expense of the environment. If the remanufacturer adopts full remanufacturing, it can be inferred to a certain extent that the market is relatively mature and consumer acceptance is high. At this stage, innovation can more rapidly achieve the goal of reducing environmental impact.

6. Conclusions

6.1. Summary and Managerial Implications

Product innovation for remanufacturing can significantly reduce material consumption and environmental pollution in remanufacturing, thereby promoting the growth of the remanufacturing industry. However, product innovation for remanufacturing leads to a series of changes in third-party remanufacturing, influencing the decisions of both OEMs and remanufacturers. Consequently, further research is needed to explore the changes brought about by product innovation for remanufacturing. Furthermore, the emergence of diverse closed-loop supply chain structures in various third-party remanufacturing modes introduces variability in innovation decisions.

This study conducted a theoretical exploration of this issue by developing models that integrate product innovation for remanufacturing into two third-party remanufacturing scenarios, providing a novel perspective on addressing the problem. The findings are summarized as follows:

(1)

The OEM always engages in product innovation for remanufacturing in third-party remanufacturing. However, not all innovation opportunities are valuable. When the industry is highly developed, or investment efficiency is high, the optimal level of innovation is sufficiently high. Conversely, with lower levels of industry development or investment efficiency, the optimal innovation level decreases, demonstrating that more innovation is not always better. Compared to the AR mode, the OR mode offers a wider range of scenarios where product innovation for remanufacturing can generate profits for the OEM.

(2)

In both third-party remanufacturing modes, there are certain remanufacturing cost conditions for the remanufacturer to enter the market. Compared with the single market result from Zou et al. [35], this paper discovered three market entry strategies for the remanufacturer. The remanufacturing strategies of the remanufacturer are primarily influenced by consumer preferences and recycling costs. Product innovation for remanufacturing can lower the entry condition for the remanufacturer and increases the likelihood of adopting a full remanufacturing strategy. The conditions for market entry remain the same under both AR and OR modes. However, the OR mode increases the likelihood that remanufacturers will adopt a full remanufacturing strategy.

(3)

Although product innovation for remanufacturing can bring profits to the OEM and remanufacturer, it does not consistently reduce environmental impact. In the forward supply chain, increasing the level of product innovation always affects the environment. However, in the reverse supply chain, when remanufactured products are widely accepted by consumers, innovation can reduce the overall environmental impact. When a full remanufacturing strategy is adopted, the OR model results in a higher environmental impact. In addition, as the level of innovation increases, the AR model is initially more environmentally friendly, followed by the OR model becoming more eco-friendly.

These finding can offer some managerial implications for firms. First, OEMs should adjust their innovation strategies to the investment efficiency. In mature industries where R&D efficiency is high, OEMs should pursue high-level product innovation to maximize profitability. Conversely, for firms with lower capital efficiency, adopting a conservative, incremental innovation strategy is more prudent to avoid diminishing returns. Second, OEMs can guide remanufacturers toward more environmentally sustainable remanufacturing strategies by adjusting the remanufacturing mode and innovation level, thereby optimizing the environmental performance of the supply chain. Finally, addressing the challenge of low market acceptance, OEMs can employ visible product innovations (e.g., modular design) to show high quality to consumers, or leverage the brand endorsement to boost purchase intention of consumers.

6.2. Limitations and Future Research

Although this study has obtained some meaningful findings, there are still some limitations. It is assumed that the willingness of consumers to buy increases with the socially responsible behaviors of the OEM. However, the purchase intention may also be influenced by other factors, such as quality, which can be further observed. Furthermore, this study is conducted within a single cycle, while a longer cycle can better reflect the process of product innovation for remanufacturing level improvement. This presents an interesting area for future research.