The role of Fintech in circular economy practices to improve sustainability performance: a two-staged SEM-ANN approach

Abstract

Coupling the practice-based view (PBV) of firms with dynamic capabilities theory (DCT), we assess the effect of Fintech adoption (FA) on organizational sustainability performance (SP) through circular economy practices (CEP). Additionally, this research examines the moderating roles of a firm’s access to finance (AF) and absorptive capacity (AC) in the interplays between the constructs. Three hundred responses were collected from Bangladeshi manufacturing SMEs using a structured questionnaire. We examined our conceptual model using a two-staged structural equation modeling-artificial neural network (SEM-ANN) approach. The empirical findings unveiled that Fintech adoption significantly drives organizational CEP and SP and that CEP acts as a mediator between the FA and SP linkage. Furthermore, the findings also confirmed the moderating effect of AF on the FA and CEP association and the impact of AC on the CEP and SP association. Hence, this scholarship adds pivotal insights to the extant literature by establishing the roles of multiple mediators and moderators in the interplay of FA and firms’ SP. Given the paucity of primary-data-based research, this empirical study addresses the gaps in the Fintech, CE, and sustainability literature and yields crucial implications for theory and practice.

Introduction

Due to institutional pressure and environmental legislation, small, medium, or large businesses now prioritize sustainability. The small and medium-sized enterprises (SMEs) sector is an integral aspect of the global economy, in which sustainability principles garner worldwide attention (Gani et al. 2022). Governments are imposing stricter regulations to reduce the negative environmental impacts of manufacturing and related activities, which has prompted firms to adopt the concepts of sustainable production (Huang and Badurdeen 2017). Consequently, preserving ecological sustainability and social responsibility and recognizing drivers that promote SME growth are the primary challenges facing stakeholders and policymakers in the changing business landscape (Yumei et al. 2021). Given the dominance of SMEs in the global economy, they are expected to be accountable for around 70% of worldwide industrial waste pollution (Caldera et al. 2019). These organizations must manage pollution mitigation and align their company strategies with sustainable business practices (Malesios et al. 2021). Now, firms prioritize environmental sustainability due to public knowledge and consumer preference for environmentally friendly products (Rehman et al. 2021). However, many manufacturing SMEs have inefficiencies in producing goods and services (Bresciani et al. 2022b). Manufacturing SMEs confront numerous obstacles when attempting to incorporate sustainability practices (Iacovidou et al. 2021). Literature has presented several barriers, such as a lack of financial resources, a high initial capital cost, and a lack of expertise (Caldera et al. 2019; Gupta et al. 2020a; Jesus et al. 2021). Moreover, there is less consensus in the literature regarding the determinants that influence a firm’s sustainability performance (Dey et al. 2020; Pizzi et al. 2021; Rodríguez-Espíndola et al. 2022). Recent studies posit that Industry 4.0 technologies can play a crucial role in efficiently utilizing resources, which can contribute to the sustainability performance of businesses (Jabbour et al. 2020; Khan et al. 2021b; Ali et al. 2022; Tang et al. 2022). In a similar vein, we argue that I4.0 technology, like financial technology (Fintech), can be a crucial antecedent of SMEs’ sustainability performance.

Fintech refers to the most contemporary technology utilized in creative financial goods and services and is regarded as one of the most ground-breaking and inventive industries of recent years (Liu et al. 2021; Najaf et al. 2023). Although the Fintech industry is yet in its early phases, the rising trend of investments in this technology implies acceptability and confidence among stakeholders (Chen et al. 2021). Fintech uses widespread machine-to-machine connectivity and the internet of things (IoT) in the financial services industry (Huynh et al. 2020; Pizzi et al. 2021). However, the existing research has not adequately investigated the relationship between Fintech adoption (FA) and the sustainability performance of firms, especially in the context of SMEs. Prior literature on FA examines the impact of Fintech on firms’ financial performance (Liu et al. 2021), renewable energy use (Croutzet and Dabbous 2021), access to finance (Abbasi et al. 2021), and agricultural sustainability (Anshari et al. 2019). However, only a few empirical research evaluates the direct and indirect effects of FA on the sustainability performance of firms (Pizzi et al. 2021; Vergara and Agudo 2021). In addition, environmental management scholars recommend exploring the presence of different mediators in the relationship between FA and sustainability performance. To address this research gap, this study also evaluates how circular economy practices (CEP) mediate the relationship between FA and SP of SMEs. “The circular economy is a concept that encourages the most efficient use of resources by employing recycling, reuse, and recovery techniques” (Tang et al. 2022, p. 2). Under Industry 4.0, the CE is related to various technological breakthroughs, including Fintech, blockchain technology, artificial intelligence, e-commerce, and big data analytics (Dalenogare et al. 2018). These technological advancements improve the current infrastructure by safeguarding sustainability objectives through innovations (Huynh et al. 2020). Thus, the inclusion of CEP is of utmost relevance, as it can guide SME managers on how to efficiently employ financial resources to enhance the sustainability performance of businesses. Additionally, prior studies have established access to finance (AF) as a significant barrier to CE (Jesus et al. 2021) and absorptive capacity (AC) as a crucial predictor of firms’ SP (Delmas et al. 2011; Aboelmaged and Hashem 2019). However, evidence of their moderating effects on the FA-CEP and CEP-SP linkages is scant in the literature.

Despite advancements in the literature on technology adoption and environmental sustainability, the linkage between organizations’ technology adoption and sustainability performance remains understudied and deserves deeper investigation (Pizzi et al. 2021; Yan et al. 2022). The environmental sustainability literature has primarily focused on strategic drivers, including ecological strategy (Rehman et al. 2020), sustainability orientation (Rehman et al. 2022), corporate social responsibility (Kraus et al. 2020), and green practices, including green leadership (Akram et al. 2022), green innovation (Rehman et al. 2021; Bresciani et al. 2022a) that increases firms’ environmental sustainability performance. However, there is a dearth of literature on how manufacturing companies’ technology adoption can facilitate the business’s transition to a CE model and improve environmental sustainability (Pizzi et al. 2021; Ramakrishna 2022). In the context of Fintech, most of the FA literature comprises reviews and case studies; thus, Pizzi et al. (2021) call on researchers to examine the influence of FA on firms’ SP through primary-data-based empirical analysis. Another gap in the extant literature is that it does not offer a good understanding of the crucial role of financing in implementing green practices such as CEP and boosting firms’ green competitive advantage (Guang-Wen and Siddik 2022). Recent scholarships have advocated that better AF enables SMEs to adopt green supply chain management and implement CEP (Gonçalves et al. 2022). Thus, bridging the aforementioned research gaps, this scholarship intends to address the following research question: What intricate associations exist between SMEs’ FA, AF, CEP, AC, and sustainability performance?

We draw insights from the practice-based view (PBV) of firms (Bromiley and Rau 2014) to explore the role of Fintech adoption and circular economy practices in enhancing SMEs’ sustainability performance. The PBV perspective offers an appropriate theoretical lens for assessing firms’ sustainability performance as it provides insights into how imitable and transferable practices can improve organizational performance (Khan et al. 2022). The PBV attempts to explain performance in part through “imitable activities or practices, often in the public domain, amenable to transfer across firms” (Bromiley and Rau 2014, p. 1249). We believe that technological practices such as Fintech implementation and circular economy practices can be imitable and transferable from one firm to another and can be utilized to improve environmental and economic performance. This study also underpins the dynamic capability theory (DCT) (Teece et al. 1997b) to conceptualize AC as a critical DC of firms that helps explore and exploit necessary knowledge to improve firms’ sustainability performance. The DC perspective of sensing, seizing, and transforming capabilities offers the fundamental framework to develop individual and firm-level absorptive capacities (Teece 2014). Thus, we argue that the DCT provides a sound theoretical lens to assess how SMEs can attain superior sustainability performance through exploiting absorptive capacities in their circular economy practices.

This study attempts to examine the interconnections between FA, AF, CEP, AC, and SP and, in doing so, makes several significant theoretical contributions. First, this scholarship adds to the scant research on Fintech adoption and its influence on corporate SP. Our research is one of the first to assess Fintech innovation as a critical antecedent of the CEP and SP of manufacturing SMEs. Second, we evaluate the CEP’s ability to mediate interactions between FA, SP, and AF and SP. Thus, outline the mechanism through which financial factors such as financial technology and capital might impact the environmental sustainability of SMEs. Third, to the researchers’ knowledge and following a review of research databases, no prior research has empirically explored the impact of FA and AF on these two outcomes in the setting of SMEs. However, research indicates that a lack of technology and finance poses significant obstacles to SMEs’ adoption of CEP and sustainability initiatives (Ghisetti and Montresor 2020; Sohal et al. 2022). Fourth, our research enhances the understanding of the crucial moderating role of AF and AC in achieving CE and sustainability goals, paving the way for future research. Finally, this study draws insights from two theoretical approaches (i.e., PBV theory and dynamic capabilities theory). This study expands the research on Fintech adoption and businesses’ environmental outcomes (CEP and SP) by integrating the two theories and examining the distinct mechanisms of access to finance and absorptive capacity. Hence, this study contributes to the theoretical development of PBV and DC theories concerning adopting financial technology and how it boosts CEP and SP through better AF and AC. Moreover, the results of this research provide significant insights for Bangladeshi manufacturing SME entrepreneurs on how to leverage I4.0 technologies like Fintech to achieve superior environmental sustainability.

The subsequent sections of the study include a discussion of the theoretical framework and the development of hypotheses. The part on hypothesis development is followed by the section on research methods, which includes sample and data collection methods, measurements, and analysis techniques. Following the explanation of the methodology, the findings of the analysis are presented and interpreted in the subsequent section, followed by a discussion. The discussion section compares the results to the existing literature and addresses the theoretical, managerial, and policy implications. The final section covers the limitations and future directions of the research.

Theoretical framework

Practice-based view

The practice-based view (PBV) is concerned with elucidating the influence of imitable or publicly accessible organizational practices on firm performance. Bromiley and Rau (2014, p. 1249) define practices as “a defined activity or set of activities that a variety of firms might execute.” The PBV provides a novel and distinct perspective on strategy literature, supplementing existing perspectives like the resource-based view (RBV) of organizations. However, the PBV centers on “imitable activities or practices, often in the public domain, amenable to transfer across firms” Bromiley and Rau (2014, p. 1249). Unlike the RBV, which describes an organization’s competitive edge in terms of unique resources, the PBV describes organizational performance in terms of generally recognized and attainable activities that businesses may engage in (Bromiley and Rau 2016). According to Hitt et al. (2016), the original RBV emphasized macro-level outcomes. As the theory has progressed, scholars have proposed that its outcome variable become increasingly focused on micro-level (Foss 2011) or mid-level outputs (Sirmon et al. 2011), including such innovation and sustainability or environmental performance. Unlike frameworks like the RBV, which concentrates on competitive advantage, the PBV allows for the study of sustainability practices on the full spectrum of performance outcomes (Carter et al. 2017; Betts et al. 2018).

Given that sustainability is a societal objective that must be accomplished, there is a compelling case for SMEs and their managers to enact and disseminate sustainability-related processes that are easily imitable and transferrable rather than pursuing a competitive edge through unique and inimitable resources (Brömer et al. 2019). Despite their significance, there is limited literature that utilized the PVB framework. As a few notable exemptions, Khan et al. (2021c, 2022) and Tang et al. (2022) used the PBV as a theoretical lens to explore the effect of CEP on the SP of organizations. Similarly, this study supports the PBV approach, which suggests that easy-to-imitate sustainability practices (such as circular economy practices) will strongly influence SMEs’ sustainability performance. The PBV paradigm is a valuable theoretical lens for analyzing the performance outcomes of SMEs since it provides insights into the shifts in company performance resulting from the adoption of imitable and transferable practices (Tiwari et al. 2020). Prior research identifies circular economy and financial technology deployment as the critical practices that may be included in green supply chain management to boost the sustainability performance of businesses (Khan et al. 2022). Regarding CE for SMEs, the PBV offers a supportive stance since it asserts that CE practices may be transferred across firms without the need for robust isolation mechanisms (Tang et al. 2022). SMEs can apply the 3R (reduce, reuse, and recycle) principles of CE, for instance, by emulating the operations of publicly listed large manufacturing enterprises. In addition, the PBV gives a theoretical foundation for enterprises’ adoption of innovative technologies (Rivera and Cox 2014; Dubey et al. 2022). A practice viewpoint on technology adoption emphasizes duality, dynamics, reciprocal relationships, connectedness, and dispersed agency to guide technology adoption theory and practice. Understanding these concepts facilitates the successful adoption of technologies, such as Fintech, that influence organizational and social performance, such as economic prosperity, enhanced productivity, entrepreneurship, sustainability, and other economic, social, and environmental improvements (Kannan and Perez-Aleman 2022). Hence, we draw upon firms’ PBV to investigate the role of SMEs’ technology adoption and CE practices in enhancing sustainability performance.

Dynamic capabilities theory

The dynamic capability (DC) perspective is a notion that refers to an organization’s capacity to react to and conform to dynamic environments by integrating and reorganizing available resources effectively (Teece et al. 1997a). DCs increase growth and profitability by boosting the organization’s adaptability to a complex and turbulent environment (Ali et al. 2022). DCs are inherently useful for adapting to competitive market situations. DC theory (DCT) is regarded as a process of obtaining and developing new skills that contribute to the enhanced performance of an organization (Gupta et al. 2020b). In the literature, several manifestations of the DC paradigm have been presented. Most prevalent is Teece’s (2014) framework, which comprises three primary components: “sensing,” “seizing,” and “transforming.” Sensing is an organization’s capacity to uncover, assess, and evaluate technological options that may be leveraged to address the organization’s customer needs and strategic goals (Wamba et al. 2020). Seizing capacity refers to utilizing appropriate methods and resources to capitalize on recognized opportunities to capture business value (Cheng et al. 2022). The transforming capability, also known as the reconfiguration capability, covers all operational processes that reconfigure resource sets and common capabilities to respond rapidly to market movements (Hendry et al. 2019). This study adopts the DC viewpoint as its theoretical framework to establish absorptive capacity as a dynamic organizational capability that can enhance CEP leading to improved sustainability performance. Since the PBV does not offer a valuable framework for understanding the role of AC in impacting SMEs’ sustainability performance, we integrated the DC theory into our study.

As per Volberda et al. (2010), the origins of AC can be linked to theories of dynamic capabilities that emphasize the cumulative character of knowledge. The application of AC as a dynamic capability enables a company to increase its international knowledge base and capitalize on external competence sources. Based on dynamic capabilities, AC covers three interrelated learning processes: exploration, assimilation, and exploitation, which can improve business outcomes (Sáenz et al. 2014). Literature on AC viewed AC as a dynamic capability related to generating and using the knowledge that boosts a company’s ability to obtain and maintain a competitive advantage (Zahra and George 2002). AC is considered a dynamic capacity inherent in a company’s routines and processes, making it possible to examine the stocks and flows of a company’s knowledge and link them to developing and maintaining competitive advantage. Incorporating the DC theory, we investigate the function of AC as a dynamic organizational capability in enhancing SMEs’ sustainability performance.

Hypotheses development

Fintech adoption and sustainability performance

Financial technology (Fintech) is a crucial topic and application that should prompt research in many industries since digital connectivity is critical for sustainable performance and productivity (Hammadi and Nobanee 2019). Leong and Sung (2018, p. 75) define Fintech as “any innovative ideas that improve financial service processes by proposing technology solutions according to different business situations, while the ideas could also lead to new business models or even new businesses.” Extant literature suggests that Fintech, as an example of a sector that has evolved because of Industry 4.0, has the potential to help SMEs shift towards a more sustainable business model (Pizzi et al. 2021). Fintech uses innovation and advanced technology to enhance, develop, and automate financial services to aid and support enterprises, shareholders, and customers handle their financial activities using sophisticated applications and software (Vergara and Agudo 2021). Fintech’s influence on lending is indicated by the development of various routes to obtaining monetary funds (Pizzi et al. 2021). A case in point is peer-to-peer (p2p) financing, wherein SMEs and shareholders can lend or borrow funds to develop social or sustainable initiatives (Mild et al. 2015). Moro-Visconti et al. (2020) argue that Fintech services offer small businesses an option for sustainable financing through microfinance and crowdfunding. Recent studies on the linkage between digital finance and environmental sustainability have placed sustainable digital finance and Fintech at the focus of their investigation (Yan et al. 2022).

Digitalization has the potential to substantially expedite the deployment of energy efficiency and renewable energy solutions since accelerated automation and advanced data analytics lead to reduced energy consumption (Liu et al. 2022). Fintech greatly impacts social, environmental, and ecological benefits in promoting fund usage for energy efficiency in this setting (Deng et al. 2019). The adoption of Fintech has a major impact on a firm’s consumption, savings, and investment decisions, and hence has become a driving factor for renewable energy usage (Croutzet and Dabbous 2021). Muganyi et al. (2021) highlight that a firm’s utilization of Fintech may influence its environmental sustainability performance by increasing environmental investments, reducing carbon emissions, and enhancing resource efficiency. Moreover, I4.0 technologies can be utilized to improve the social sustainability performance of an organization. (Tasleem et al. 2019) argue that businesses use their commitments, duties, and activities to meet stakeholders’ social expectations and create value for society by implementing various technologies. Financial technologies can promote investments in corporate social responsibility (Liu et al. 2021), enhancing firms’ social sustainability performance. Guang-Wen and Siddik (2022) report that by adopting Fintech, firms can achieve superior environmental performance through green financing and innovation. Since several prior studies have established the linkage between firms’ technology adoption and sustainability performance, we draw insights from the PBV of firms to posit that Fintech adoption, if integrated as a practice throughout the supply chain, can drive the sustainability performance of firms. Thus, we hypothesize that:

• H1: Fintech adoption is positively related to firms’ sustainability performance.

Fintech adoption and circular economy practices

The circular economy (CE) is a current buzzword, and as such, it receives a great deal of rhetorical attention. The meanings of trending concepts tend to diffuse, and scholars assert that the CE concept is no exception. Drawing upon 144 CE literature, Kirchherr et al. (2017, p. 229) define CE “as an economic system that replaces the ‘end-of-life’ concept with reducing, alternatively reusing, recycling, and recovering materials in production/distribution and consumption processes.” The CE principles are implemented comprehensively from resource exploitation to component and product production, consumer usage, and a cascade of sharing, repair, reuse, redistribution, and refurbishment activities (Pizzi et al. 2021). CE necessitates the restoration, renewal, and upheaval of economic systems, posing profound challenges to the evolutionary processes of business practices (de Sousa Jabbour et al. 2019). Digitalization is regarded as one of CE’s catalysts thanks to its capacity to develop insights and information into assets and goods (Antikainen et al. 2018). Blockchain technology and financial technologies are considered critical drivers of the circular economy, as technologies tied to Industry 4.0 enhance resource utilization and productivity (Kristoffersen et al. 2020). Drawing on the PBV of firms (Tang et al. 2022), we posit that Fintech may play a pivotal role in ameliorating CEP in SMEs. Fintech intends to contribute to the shift from linear to CE business models (CEBMs) by enabling SMEs to access technologies like mobile payment platforms, IoT, and artificial intelligence necessary to attain the strategic flexibility required for CEBMs (Rialti et al. 2020). Consequently, CEBMs enabled by Fintech can position citizens at the heart of environmentally and sustainably essential decisions. Pizzi et al. (2021) highlight how Fintech, a sector that emerged as a result of the impact of Industry 4.0, could assist SMEs in implementing CEP in terms of value proposition, value generation, and value delivery. Fintech can facilitate the adoption of CEP by offering both sell-side and buy-side payment solutions that can be tailored. Prior research has also established that peer-to-peer lending could facilitate the growth of CEP among firms (Fischer and Pascucci 2017). Moreover, Fintech solutions promote information disclosures, risk assessments, financing and investor-matching, and insurance services, thereby expediting an organization-wide shift to CE and carbon neutrality (Ramakrishna 2022). Hence, we postulate that:

• H2: Fintech adoption is positively related to firms’ circular economy practices.

Access to finance and circular economy practices

Access to finance (AF) has evolved as a major obstacle for companies utilizing the CE business model in practice and scholarly research. Extant literature identifies the acquisition of (external) capital as a critical and well-known barrier to CEP (Iacovidou et al. 2021; Jesus et al. 2021), particularly for SMEs and young enterprises (Ghisetti and Montresor 2020; Toxopeus et al. 2021). Informational obscurity between the organization and its potential financiers, moral hazard concerns, and high transaction costs are the root drivers of credit constraints (see Cosh et al. 2009). Problems inherent to intangible R & D investments, lack of collateral, and a track record exacerbate these limitations (Mina et al. 2013; Brancati 2015). Jesus et al. (2021) highlight that a lack of financing hinders CEP since it requires significant investments in production technologies with uncertain returns. Moreover, for a myriad of reasons, inadequate access to funding tends to be a greater obstacle to the adoption of CEP among SMEs than among large enterprises (Ghisetti and Montresor 2020). First, SMEs are usually more financially limited than large corporations (Lee et al. 2015), making them more susceptible to the financing costs of implementing the tracking and development operations that CEP usually entails (Toxopeus et al. 2021). Second, the initial costs and delayed return seem to be more relevant for SMEs than for large corporations, given their greater sensitivity to the additional costs of resource efficiency research (Ghisetti and Montresor 2020). Third, the question of the lower residual value of CE resources hinders the SME’s access to bank funding to a larger extent than it does for large corporations due to their relative disadvantage when it comes to available collateral (Hyz 2011). Finally, yet importantly, external financing proves to be more challenging for SMEs given the widespread shortage of manpower and management processes in analyzing and exploiting potential resources (Beck and Demirguc-Kunt 2006). Drawing on these arguments, we posit that greater AF would facilitate investment in SMEs’ circular economy practices. Therefore, we hypothesize:

• H3a: Access to finance is positively related to firms’ circular economy practices.

Access to finance and sustainability performance

Businesses make profitable investments (i.e., those with a positive net present value) to improve performance and gain a long-term competitive advantage. Nevertheless, the capacity to fund such strategic initiatives is closely linked to each firm’s particular capital restrictions (Cheng et al. 2014). This challenge could be “due to credit constraints or inability to borrow, inability to issue equity, dependence on bank loans, or illiquidity of assets” (Lamont et al. 2001, p. 529). According to Kahupi et al. (2021), investors’ skepticism about sustainability-based innovations owing to their costs, yields, and inherent risks impedes the financing of new sustainable firms. Access to external funding is crucial for the seamless functioning of a firm’s operations and for eco-friendly and sustainable practices (Knight et al. 2019). If business owners and shareholders consider the added benefits of sustainability in their overall business strategy instead of focusing exclusively on financial growth and profitability, financial hurdles can be minimized as the most significant barrier to sustainability for firms (Fotiadis et al. 2013). In a similar vein, Caldera et al. (2019) argue that one of the greatest challenges to the effective implementation of green and sustainable business practices in lean SMEs is financial constraint due to the absence of immediate quantifiable benefits, high capital costs, and declining sales resulting from the premium price integrated in green products. Ullah et al. (2021) contend that greater access to internal and external financing strongly influences different dimensions of a firm’s sustainability performance. Likewise, this study also proposes that access to finance has a positive linkage with SMEs’ sustainability performance and hypothesizes that:

• H3b: Access to finance is positively related to firms’ sustainability performance.

Circular economy practices and sustainability performance

The idea of a CE is enticing, has increased government and industry awareness and readiness to act, and has been found to help the achievement of the sustainable development goals (Schroeder et al. 2019). At the firm level, CEP incorporates environmental protection regulations on reduction, reuse, and recycling (3Rs) with a focus on meeting environmental and economic performance objectives (Zhu et al. 2010). CE is regarded as a major contributor to sustainability (Geissdoerfer et al. 2017; Corona et al. 2019; Rodríguez-Espíndola et al. 2022). While the former focuses on reducing inputs, waste, and emission levels, the latter has broader open-ended objectives dependent on the stakeholders and their respective interests (Geissdoerfer et al. 2017). CE is a business paradigm that necessitates novel methods of thinking and conducting business (Bocken et al. 2016). It is a restoring and regenerating model (Rodríguez-Espíndola et al. 2022) that aims to increase production and consumption efficiency by implementing the 3Rs (Ghisellini et al. 2016). Resource circularity, natural resource utilization, and product longevity are all results of CE-based manufacturing systems connected with sustainable operations (Gupta et al. 2019; Bai et al. 2020). CE evolved as a revolutionary paradigm that improves enterprises’ economic, environmental, and social aspects to drive society toward greater sustainability through the cooperation of all stakeholders (Dey et al. 2020). Several CE scholars and practitioners argue that environmental and economic sustainability elements are implicitly incorporated within CE (Calisto Friant et al. 2020). In a qualitative study of 155 enterprises in Italy and the Netherlands, Walker et al. (2022) observed that most of the leading firms engaging in CE regard CEP as a substantial tool for achieving superior sustainability performance, notably in the environmental domain. Hence, we argue that:

• H4: Circular economy practices are positively related to firms’ sustainability performance.

Absorptive capacity and sustainability performance

Extant literature defines absorptive capacity (AC) as a “dynamic capability that purposefully creates, extends, and modifies a firm’s resource base and as a higher-order capability that enables the development of competences and capabilities” (Dzhengiz and Niesten 2020, p. 885). Sáenz et al. (2014) note that AC’s role is to fortify relationships between buyers and suppliers for sustained business success. Aboelmaged and Hashem (2019) argue that AC is an effective predictor of sustainable capacities and the adoption of eco-friendly innovations. AC contributes to the formation of proactive environmental policies (Delmas et al. 2011), and prior research indicates that its integration into a strategic plan can boost capability-based performance (McAdam et al. 2010). Hofmann et al. (2012) establish a connection between sustainability practices and underlying capabilities, recommending that businesses develop certain capabilities before engaging in sustainability projects. Sustainability standards are one method for enhancing SP, which itself depends on AC (Sáenz et al. 2014); its application necessitates alterations to organizational structure, procedures, and values and hence depends on effective organizational learning.

Dzhengiz and Niesten (2020) establish AC as a multidimensional DC in which two dimensions — recognizing the value of external knowledge and knowledge acquisition — have a positive linkage with the development of environmental competences and other dimensions — knowledge assimilation, transformation, and exploitation — have a positive association with the development of ecological competences. Thus, AC can help firms address barriers to an environmentally sustainable philosophy and encourage the development of new capacities required to reconfigure the firm’s processes and products to lessen their negative environmental consequences (Delmas et al. 2011). AC contributes to the enforcement of sustainability practices since a successful execution requires a mix of knowledge from multiple sources, which often lie beyond organizational boundaries (McWilliams and Siegel 2001; Delmas et al. 2011). Delmas et al. (2011) stress that knowledge acquisition, integration, translation, and exploitation are crucial for developing and adopting sustainability practices. Therefore, consistent with the theoretical rationale and empirical evidence, we contend that AC as a critical dynamic capability significantly drives the SP of firms and hypothesize that:

• H5: Absorptive capacity is positively related to firms’ sustainability performance.

Mediating effects of circular economy practices

I4.0 technologies can be incorporated into value chains by acquiring and constantly exchanging data to offer real-time information regarding equipment, manufacturing, operations, and material movements; this assists management in tracking, monitoring, and making long-term decisions concerning post-consumption good recoveries (Khan et al. 2021a). Such recovery-based techniques substitute the typical linear “take, make, use, and dispose” approach with a circular one that provides social, economic, and environmental advantages to businesses and supply chains (Geissdoerfer et al. 2017). Some researchers hypothesized that CEBMs would be the most viable operational vehicles for building sustainable business models in I4.0 contexts (Bressanelli et al. 2018; Khan et al. 2021a). CE is a paradigm that aspires to the sustainable use of resources (McDowall et al. 2017), requiring a transition from the linear system to the circular system of “reduce, reuse, recycle, recover, remanufacture, and redesign” (Jabbour et al. 2020). Thus, integrating I4.0 technologies can improve a firm’s CEP, which, if successfully managed, can lead to superior sustainability performance. This scholarship, therefore, expects CEP to mediate the relationship between an SME’s adoption of Fintech and its sustainability performance. There are a few grounds for this expectation. For instance, FA may not directly impact an organization’s SP, but it may be used to restructure business models into circular and sustainable ones (Pizzi et al. 2021). Fintech seeks to assist in the shift from linear to CEBMs by offering SMEs access to technologies such as mobile payment solutions, IoT, and artificial intelligence that are essential for ensuring the dynamic capabilities needed for CEBMs (Rialti et al. 2020). Prior studies, on the other hand, have established CEP as a substantial contributor to the sustainability of businesses (Geissdoerfer et al. 2017; Rodríguez-Espíndola et al. 2022). Hence, we contend that FA can support SMEs in deploying CEP regarding value proposition, value generation, and value delivery, which may result in superior SP. Based on this rationale, we hypothesize that:

• H6: Circular economy practices mediate the relationship between Fintech adoption and sustainability performance.

Furthermore, financial resources are critical for firms’ sustainability performance. However, we argue that the association between AF and a firm’s SP is rather indirect. There could be different intervening factors like green financing, circular economy practices, and green supply chain management that mediate the impact of AF on firms’ SP. Lack of financing is one of the most crucial barriers for CEP, particularly for SMEs (Ghisetti and Montresor 2020). The shift to CE necessitates the assessment of the financial rewards attached to it, prompting a rethinking of the economic connections among the several stakeholders dealing with the organization (Geissdoerfer et al. 2017). This is especially pertinent for SMEs, which face greater financial restrictions than large corporations. Subsequently, SMEs’ implementation of sustainable practices is frequently hindered by the existence of financial hurdles (Pizzi et al. 2021). Thus, we posit that with greater access to capital, SMEs can invest in sustainable operations and circular economy practices, which would ameliorate firms’ sustainability performance.

• H7: Circular economy practices mediate the relationship between access to finance and sustainability performance.

Moderating effects of access to finance

Prior studies argue that adopting I4.0 technologies substantially affects developing CEBMs. However, one of the significant barriers to the transition toward CE is the lack of financial resources (Jesus et al. 2021). Technology implementations are not adequate for an organizational shift to CE models. We argue that access to finance can be influential in enhancing the impacts of different drivers of CEP in a firm. Moreover, the CE literature in the SME context has established that SMEs are often demotivated to invest in CEP because of financial barriers (Ghisetti and Montresor 2020; Toxopeus et al. 2021). The lack of funding and tax rebates hinders investments in the circular economy, which has high initial costs and slow return growth (Kirchherr et al. 2018). Thus, we argue that with better AF, SMEs can effectively utilize financial technologies that promote circular economy practices. Based on this argument, we hypothesize that:

• H8: Access to finance moderates the relationship between Fintech adoption and circular economy practices.

Moderating effects of absorptive capacity

As promoting the shift to a CE demands considerable changes in the processes and activities of businesses (Marrucci et al. 2022) and might bring significant competitive benefits (Gusmerotti et al. 2019), there are crucial interaction points between absorptive capacity and circular economy. Therefore, AC may be pivotal in an organization’s efforts to achieve sustainability and a circular economy. AC is “a dynamic capability that influences the firm’s ability to create and deploy the knowledge necessary to build other organizational capabilities” (Abareshi and Molla 2013, p. 211). Prior research on sustainability contends that CE and AC are crucial determinants of organizations’ sustainable performance (Delmas et al. 2011; Walker et al. 2022). Consequently, we hypothesize that AC may increase an organization’s tendency to adopt circular business models (Marrucci et al. 2022), which strongly impacts SMEs’ SP.

• H9: Absorptive capacity moderates the relationship between circular economy practices and sustainability performance.

Research methods

Study context

A survey of Bangladeshi manufacturing SMEs was conducted to evaluate the suggested hypotheses. These SMEs were selected because they operate in a dynamic and competitive market that necessitates strategic approaches for growth and success, such as the latest technology adoption and CEP. Regarding country selection, we have chosen Bangladesh as a context for our empirical study for diverse reasons (Holgersson 2013). SMEs are the backbone of the Bangladeshi economy, employing 7.8 million people directly while supporting 31.2 million (LightCastle Analytics Wing 2020). SMEs contribute around 25% of the nation’s GDP and have the potential to contribute much more. Moreover, as a developing country, Bangladesh is making considerable advances in meeting the United Nations’ sustainable development goals (SDGs) (Rahman 2021), which has contributed to the growth of several industries, including SMEs. However, Bangladesh’s vulnerable SMEs took the brunt of the recent COVID-19 pandemic. Many businesses were permanently shut down, while others suffered financial losses.

In addition, financial obstacles frequently prevent these SMEs from implementing sustainable practices. Prior research on the adoption of CE and sustainability actions by SMEs has also revealed that SMEs need financial incentives to compensate for the opportunity costs associated with reconfiguring their operations. In this regard, I4.0 technologies, such as Fintech platforms, facilitate SMEs’ implementation of CE and sustainability activities by offering alternative funding channels (Pizzi et al. 2021). Initially, Bangladeshi SMEs resisted embracing sustainable practices and adhering to environmental legislation. However, as international organizations finance SMEs to comply with environmental standards, the scenario is progressively improving. International development organizations have funded Bangladeshi manufacturing SMEs to enhance worker safety, green manufacturing technologies, and environmental considerations (Islam 2021). In Bangladesh, Agence Francaise de Développement (AFD) has contributed 50 million euros to green SME investment (Islam 2021). The Sustainable Enterprise Project (SEP) of the World Bank is directly assisting 40,000 Bangladeshi small and medium-sized enterprises (SMEs) in implementing environmental initiatives and diversifying their portfolios incorporating ecological stewardship, waste and emissions control, and improved workplace health and safety (Yoshijima et al. 2020). These small businesses continually adopt technological advancements to achieve environmental sustainability (The Financial Express 2022). Bangladeshi SMEs are using Fintech platforms that offer digital lending and collaborative credit scoring for SMEs (The Business Standard 2022). Hence, it is crucial to investigate how adopting Fintechs may enable Bangladeshi SMEs to transit to a CE business model and attain superior sustainability performance.

Data and sample

A self-administered questionnaire was constructed to obtain information on the impact of FA, AF, and AC on the CEP and SP of Bangladeshi SMEs. We pre-tested the questionnaire by two academics and four SME managers to ensure the validity of the survey instrument. The primary purpose of the pre-test was to identify questions that respondents could find unclear. Participants’ feedback suggested that a few items were ambiguous; therefore, we simplified those to improve precision. As a result of input from the pilot survey, minor adjustments were made to the survey questions. The survey data was collected from SME managers with a thorough understanding of their firm’s operations and performance. We employed simple random sampling to choose the respondents. The questionnaire was distributed to 410 SMEs along with a cover letter describing the study’s aims and emphasizing that participation was voluntary. Furthermore, respondents were informed that their responses would be kept entirely anonymous and used solely for academic research. Following a reminder, 300 usable and complete surveys were returned, yielding a response rate of 73.17%. This study’s data was collected between July 2021 and January 2022.

Measures

We assessed all the model’s hypotheses using several questionnaire items. All indicators were extracted from formerly published studies. We modified some indicators to fit the setting of the study. The exogenous variables were quantified using a five-point Likert scale. Prior to the primary inquiry, the instrument was created based on the method outlined by Mishra et al. (2016) and Hair et al. (2012). Three SME experts were provided with the complete set of study constructs. On a three-point Likert scale, participants were then asked to score the extent to which they considered these questions properly reflected their respective concepts: 3 indicates a high degree of accuracy, 2 suggests moderate accuracy, and 1 indicates no accuracy. The survey instrument included all items that were evaluated “three” by at least two professionals and were not graded “one” by any of the professionals.

This study’s measuring instruments were derived from previous literature on the topic. There is a dearth of literature on adopting Fintech at the firm level, particularly in the context of SMEs. Therefore, seven FA items were extracted from a study on banks’ adoption of Fintech by Dwivedi et al. (2021), modified according to the context of this research, and measured on a 5-point Likert scale. Respondents were asked if using Fintech has promoted innovation, provided new channels, complied with the country’s legislation, and so on. Next, access to finance was incorporated in this study as a moderator between FA and CEP. Five items were derived from the study of Wasiuzzaman (2019) to measure the AF construct. The participants were asked about the cost, availability, range of financial services, and the size of the loans available to them. Next, the mediator, CEP, was measured using 4 items adopted from Zeng et al. (2017) and Zhu et al. (2005). We measured firms’ absorptive capacity by deriving 4 items from Limaj and Bernroider (2019). Finally, we measured the sustainability performance of the SMEs using 6 items from Zhu et al. (2008); Paulraj 2011; Sajan et al. (2017). Fig. 1 represents the conceptual model that illustrates the hypothesized relationships among the constructs.

Fig. 1.

Conceptual model

Statistical analysis techniques

In the initial stage of the study, the proposed conceptual model was evaluated using the partial least squares structural equation modeling (PLS-SEM) technique utilizing SmartPLS software version 3.3.3. PLS- SEM is appropriate due to two factors. First, compared to covariance-based structural equation modeling (CB-SEM), PLS-SEM is preferable for projecting complicated models and developing theories (Leong et al. 2020), which is the focus of this study. In addition, PLS-SEM imposes fewer limitations on sample size and non-normal distributions (Leong et al. 2013). Given that the Kolmogorov–Smirnov test yielded p-values less than 0.001 for each item, the analysis concluded that the data are not multivariate normal. PLS-SEM was considered more appropriate for this study than CB–SEM (Islam et al. 2019). In addition, the PLS literature strongly recommends G*power analysis for calculating the proper sample size (Hair Jr. et al. 2016). Thus, we evaluated the sample size with the G*Power 3.1.9.7 software, an update of earlier versions. Our sample size of 300 had sufficient statistical power, exceeding the minimum sample size of 118 determined by G*power, with a power level of 0.80, 10 predictors, an alpha value of 0.05, and an effect size of 0.15. Since PLS-SEM can only detect linear associations, the artificial neural network (ANN) approach was used in the second stage of analysis to assess the relevance of the predictors by identifying the complicated linear and non-linear associations among the constructs. As non-linear interactions exist between the variables, we further performed ANN analysis using the significant predictors from the SEM analysis to rank their normalized importance (Yan et al. 2022). The graphical representation of the research methodology is illustrated in Fig. 2.

Fig. 2.

Flowchart of research materials and methods

Results

Common method bias

There is a risk of common method bias (CMB) due to the cross-sectional nature of data. Following the methods of Leong et al. (2020), we employed both statistical and procedural approaches. Regarding the procedural steps, participants were assured of their responses’ anonymity and confidentiality. Moreover, simple sentences were employed, and uncommon jargon was eliminated from the questionnaire to prevent misunderstandings. Regarding statistical solutions, Harman’s single-factor test was undertaken according to the criteria of Podsakoff et al. (2003), and the results indicated that an exploratory factor analysis including all variables yielded a single component accounting for 46.5% of the total variance. Additionally, the inter-construct correlations were evaluated to determine if they exceeded the 0.90 thresholds (Bagozzi et al. 1991). Given that the highest value of inter-construct correlation was 0.766, it can be concluded that CMB is not a severe concern in the data used.

First stage: SEM

Measurement model

First, we assessed the reliability and validity of all the measures employed in this research. The outcomes are shown in Table 1 below. The average variance extracted (AVE), Cronbach’s alpha (CA), and composite reliability (CR) are displayed in Table 2 as metrics of reliability and validity. The values indicated that the validity and reliability meet the established standards. All constructs exceeded the acceptable value of 0.70 for CR (Wasko and Faraj 2005). CA values for each construct were also above the 0.70 thresholds, indicating excellent construct reliability as per Hair et al. (2016). The principle underlying convergent validity (CV) is that the theoretical foundation for associated measures is statistically related. AVE values greater than 0.5 were utilized to diagnose CV (Fornell and Larcker 1981). The CV is further supported by the fact that each model item has a considerable and significant standard loading on its target construct. Except for the AF1 item (0.676) of the access to finance factor, all outer loadings were greater than 0.70. However, in line with the guidance of Hair et al. (2016), it was retained for further investigation as the loading was statistically significant, and its removal did not improve the CR of the respective factor. The AVE values of our model constructs ranging from 0.562 to 0.727 corroborate the CV of the measures (Table 1).

Table 1.

Measurement model results

Constructs	Items	Factor loadings	Alpha	CR	AVE
Fintech adoption	FA1	0.763	0.894	0.917	0.612
	FA2	0.790
	FA3	0.785
	FA4	0.789
	FA5	0.813
	FA6	0.775
	FA7	0.760
Access to finance	AF1	0.676	0.809	0.864	0.562
	AF2	0.835
	AF3	0.791
	AF4	0.703
	AF5	0.730
Circular economy practices	CEP1	0.788	0.874	0.914	0.727
	CEP2	0.860
	CEP3	0.887
	CEP4	0.872
Absorptive capacity	AC1	0.834	0.816	0.879	0.645
	AC2	0.765
	AC3	0.788
	AC4	0.823
Sustainability performance	SP1	0.835	0.898	0.921	0.662
	SP2	0.827
	SP3	0.777
	SP4	0.827
	SP5	0.819
	SP6	0.794

Alpha, Cronbach’s alpha; CR, composite reliability; AVE, average variance extracted

Table 2.

Discriminant validity

	FA	AF	CEP	AC	SP
Fintech adoption	0.782
Access to finance	0.617	0.749
Circular economy practices	0.665	0.632	0.853
Absorptive capacity	0.759	0.624	0.763	0.803
Sustainability performance	0.744	0.674	0.766	0.737	0.813

bold values on the diagonal in the correlation matrix are square roots of AVE (variance shared between the constructs and their respective measures). Off-diagonal elements below the diagonal are correlations among the constructs

Discriminant validity (DV) was then tested, employing the Fornell-Larcker criterion to empirically establish that all the constructs are distinct from one another. Table 2 displays that the square roots of the AVEs are greater than the inter-construct correlations (Fornell and Larcker 1981). To ensure acceptable DV, the diagonal values in the associated rows and columns should be substantially bigger than the off-diagonal elements (Roldán and Sánchez-Franco 2012). As illustrated in Table 3, this requirement stands true for all the constructs of our measurement model.

Table 3.

Goodness of fit index and predictive power of the model

Constructs	AVE	R2	Q2
Fintech adoption	0.612
Access to finance	0.562
Circular economy practices	0.727	0.518	0.390
Absorptive capacity	0.645
Sustainability performance	0.662	0.766	0.504
Average scores	0.642	0.642
AVE x R2	0.412
$\mathrm{GoF}=(\surd (\mathrm{AVE}\ast R^2))$	0.642

AVE, average variance extracted; GoF, goodness of fit

Structural model

The dataset was examined for potential multicollinearity issues before analyzing the structural model. The inner variance inflation factor (VIF) scores were checked for each construct to ensure the absence of multicollinearity. According to Hair et al. (2011), a VIF score below 5 shows a lack of multicollinearity. The highest VIF value recorded was 2.603, confirming the absence of multicollinearity. After ensuring that the model does not present any multicollinearity problem, it was suitable for PLS-SEM.

We used SmartPLS 3.3.3 to assess the structural model and test the proposed hypotheses in this investigation (see Fig. 3). The bootstrapping technique was carried out with 5000 subsamples and 300 observations to assess the significance of the linkages among our model’s constructs. Since PLS does not provide overall goodness of fit metrics, R² and Q² are the principal methods for determining the predictive potential of the structural model (Wasko and Faraj 2005). Table 4 demonstrates that all R² values are greater than 0.1 (CEP = 0.518, SP = 0.766), indicating strong explanatory power of the latent variables. As a result, the capability for prediction is established (Falk and Miller 1992). Additionally, Q² proves the predictive relevance of endogenous components. A Q² value greater than 0 indicates that the model is predictively relevant. The results show that the predictive relevance of our constructs is high (CEP = 0.390, SP = 0.504; see Table 4).

Fig. 3.

Structural model

Table 4.

Results of hypotheses testing

Direct effects
Hypotheses	Coefficients	t-statistics	Remarks
H1: FA —> SP	0.126 (0.046) **	2.700	Supported
H2: FA —> CEP	0.515 (0.048) ***	10.728	Supported
H3a: AF —> CEP	0.408 (0.052) ***	7.837	Supported
H3b: AF —> SP	0.152 (0.034) ***	4.500	Supported
H4: CEP —> SP	0.245 (0.049) ***	5.055	Supported
H5: AC —> SP	0.450 (0.054) ***	8.328	Supported
Mediating effects
H6: FA —> CEP —> SP	0.126 (0.028) ***	4.612	Supported
H7: AF —> CEP —> SP	0.100 (0.024) ***	4.214	Supported
Moderating effects
H8: FA x AF —> CEP	− 0.232 (0.035) ***	6.560	Supported
H9: CEP x AC —> SP	0.135 (0.038) ***	3.591	Supported

FA, Fintech adoption; AF, access to finance; CEP, circular economy practices; AC, absorptive capacity, SP, sustainability performance; *p < 0.05, **p < 0.01, ***p < 0.001

However, Tenenhaus et al. (2005) developed the goodness of fit (GoF) index, a standardized method for assessing the model’s fit. The GoF statistic is computed by taking the square root of the product of the mean AVE and mean R² values (for endogenous variables). Wetzels et al. (2009) determined the proper cut-off values for evaluating the outcomes of the GoF estimate as follows: GoFsmall = 0.1, GoFmedium = 0.25, and GoFlarge = 0.36. The GoF value of our model is 0.642, as shown in Table 3, reflecting an excellent model fit.

After assessing the model’s multicollinearity, predictive relevance, and goodness of fit, we analyzed the direct and indirect paths between FA, AF, CEP, AC, and SP. Table 4 contains each hypothesis’s path coefficients, standard errors, and t-values.

Table 4 reports both the direct and indirect effects of the model constructs. First, we observed that FA significantly positively impacts firms’ SP (β = 0.126, t = 2.700, p = 0.0007), H1 was confirmed. Moreover, FA strongly influences firms’ circular economy practices (β = 0.515, t = 10.728, p = 0.000), supporting H2. The effects of FA on CEP were observed to be substantially larger and more significant than its effects on a firm’s SP. Next, AF has a significant and positive association with firms’ CEP (β = 0.408, t = 7.837, p = 0.000), confirming H3a. Additionally, the findings confirmed the strong linkage between AF and firms’ SP (β = 0.152, t = 4.500, p = 0.000). Hence, H3b was also accepted. Furthermore, circular economy practices were observed to have a strong positive effect on a firm’s SP (β = 0.245, t = 5.055, p = 0.000), thereby validating H4. Lastly, the direct impact of AC on SP was also confirmed (β = 0.450, t = 8.328, p = 0.000), thus supporting hypothesis H5 (see Fig. 3).

The mediation analysis revealed that CEP has a robust mediating effect on the FA-SP linkage (β = 0.126, t = 4.612, p = 0.000), supporting H6. Next, we observed that firms’ CEP also mediates the association between AF and SP (β = 0.100, t = 4.214, p = 0.000); thereby, H7 was also confirmed. To incorporate AF and AC as moderators in our SEM analysis, we created an interaction variable with standardized indicators to prevent collinearity issues and compare coefficients (Henseler and Chin 2010). The analysis demonstrates a significant but negative moderating effect of AF (β =  − 0.232, t = 6.560, p = 0.000), implying that greater access to finance mitigates the impact of FA on a firm’s SP. Thus, H8 was accepted. Finally, we observed a positive and significant moderating role of AC in the CEP-SP linkage (β = 0.135, t = 3.591, p = 0.000), indicating that better AC strengthens the relationship between a firm’s CEP and SP. Hence, all the hypotheses were supported. Figures 4 and 5 illustrate the moderating effects of AF and AC.

Fig. 4.

Moderating effect of access to finance

Fig. 5.

Moderating effect of absorptive capacity

Second stage: artificial neural network approach

Drawing insights from Liébana-Cabanillas et al. (2017), we used the significant variables of the PLS-SEM path analysis as input neurons for the ANN model in the subsequent stage (Fig. 6). The ANN analysis in this study was conducted using the SPSS neural network (NN) module. The ANN algorithm can capture both linear and nonlinear linkages and does not require a normal distribution (Leong et al. 2013). A multi-layer perceptron (MLP) utilizing the FFBP algorithm was reported in the current study. MLP has three layers: input, hidden, and output. SPSS 20.0 supported tenfold cross-validation with a data partition of 90:10 for testing and training. The hidden neurons were generated in an impulsive manner, and the sigmoid function was applied to both the hidden and output layers. According to the number of non-zero synaptic weights connected to the neural network’s hidden layer, the importance of the predictor variables was established (Leong et al. 2020). Given that only significant factors from SEM are employed in ANN models, the conceptual model can be split into two ANN models, and an illustration of one of them is portrayed in Fig. 6.

Fig. 6.

ANN model for SP output (model 1)

Model 1 (output—SP) has four inputs: FA, AF, CEP, and AC, while model 2 (output—CEP) only has two: FA and AF. To prevent model overfitting, a tenfold cross-validation approach was conducted, with 90% of the data used for network training and 10% for testing (Leong et al. 2013; Liébana-Cabanillas et al. 2017). Root mean square of error (RMSE) was used to evaluate the accuracy of NN models (Liébana-Cabanillas et al. 2017). The RMSE values for training and testing data sets (2 models, 10 times each) are shown in Table 5. RMSE is calculated by applying this formula: $RMSE=\surd \frac{\mathit{SSE}}{N}$; where SSE is the sum of square error of the training or testing data, and N is the sample size of the training or testing data. The mean values of RMSE for SP on the training dataset were 0.561, and on the testing dataset, they were 0.512. The CEP had mean values of 0.796 for training datasets and 0.746 for testing datasets, indicating that both models produced reliable estimate.

Table 5.

RMSE values of artificial neural networks

Network	Model 1		Model 2
	Output: SP		Output: CEP
	Training	Testing	Training	Testing
1	0.531	0.550	0.778	0.649
2	0.620	0.431	0.878	0.854
3	0.551	0.541	0.823	0.621
4	0.542	0.554	0.767	0.654
5	0.623	0.533	0.835	0.797
6	0.528	0.477	0.759	0.899
7	0.569	0.516	0.737	0.838
8	0.548	0.515	0.824	0.542
9	0.566	0.527	0.814	0.803
10	0.533	0.477	0.746	0.800
Mean	0.561	0.512	0.796	0.746
SD	0.035	0.039	0.046	0.119

To assess the predictive ability of each input neuron, we performed sensitivity analysis (Table 6) to derive the normalized importance of each neuron by dividing its relative importance by its maximum importance and presenting the result as a percentage (Leong et al. 2013). As per the ANN sensitivity analysis provided in Table 6, the most significant determinants of SP are absorptive capacity, followed by Fintech adoption, circular economy practices, and access to finance.

Table 6.

Sensitivity analysis

Neural network (NN)	Model 1				Model 2
	Output: SP				Output: CEP
	FA	AF	CEP	AC	FA	AF
NN (i)	0.191	0.093	0.214	0.501	0.590	0.410
NN (ii)	0.190	0.262	0.139	0.408	0.785	0.215
NN (iii)	0.200	0.195	0.178	0.427	0.571	0.429
NN (iv)	0.193	0.138	0.209	0.460	0.647	0.353
NN (v)	0.449	0.075	0.296	0.180	0.500	0.500
NN (vi)	0.182	0.121	0.225	0.472	0.631	0.369
NN (vii)	0.115	0.179	0.198	0.508	0.560	0.440
NN (viii)	0.246	0.094	0.195	0.465	0.523	0.477
NN (ix)	0.118	0.159	0.203	0.520	0.604	0.396
NN (x)	0.237	0.099	0.217	0.447	0.579	0.421
Average importance	0.212	0.141	0.208	0.439	0.599	0.401
Normalized importance (%)	49.00	33.00	47.00	100.00	100.00	69.00

Discussion

The present study’s findings indicate the relative importance of firm-level FA in influencing CE and sustainability performance. Several contributions are made to the literature on the causal relationship between FA, AF, CEP, and AC and sustainability performance. First, the finding revealed that Fintech adoption favorably affects the sustainability performance of businesses. This finding is consistent with prior studies that associated Fintech with triple-bottom-line sustainability performance (Muganyi et al. 2021; Pizzi et al. 2021). For instance, a firm’s use of Fintech may impact its environmental sustainability performance by boosting environmental investments, reducing carbon emissions, and improving resource efficiency (Muganyi et al. 2021). In addition, Fintech adoption improves peer-to-peer (p2p) financing, allowing SMEs and shareholders to lend or borrow funds to build environmental or sustainable initiatives (Mild et al. 2015; Pizzi et al. 2021). Fintech can also encourage investments in corporate social responsibility, thereby improving the social sustainability performance of SMEs (Liu et al. 2021). It is expected that in SME sector, where financial barriers are well documented, adopting financial technologies can aid businesses in obtaining funds for investing in environmental and sustainable projects. A significant linkage between FA and SP cements the PBV that argues that technological practices like Fintech adoption that can be imitated and transferred across firms lead to superior performance.

Second, our findings suggest that FA is positively associated with circular economy practices in firms. This finding is in line with previous studies which reported a strong effect of I4.0 technologies like Fintech on firms’ circular economy practices (Kristoffersen et al. 2020; Pizzi et al. 2021; Ramakrishna 2022). Recent CE literature emphasizes that Fintech enables SMEs to access technology such as mobile payment platforms, IoT, and artificial intelligence, which are essential to achieve the strategic flexibility necessary for CEBMs (Rialti et al. 2020; Pizzi et al. 2021). Fintech may facilitate the CE by augmenting financial information with product details, including material characteristics, emissions along the supply chain, guidelines for disassembly/recycling, and product profiles and images for the resale/share economy. Furthermore, Fintech advancements simplify data disclosures, risk assessments, financing, investor matching, and insurance, accelerating the system-wide shift to a circular economy (Ramakrishna 2022). Third, we observe that improved access to finance drives circular economy activities substantially. The existing literature has previously demonstrated that access to financial resources is one of the most significant obstacles to adopting CE (Iacovidou et al. 2021; Jesus et al. 2021). Furthermore, SMEs have more significant financial obstacles than large firms (Lee et al. 2015), making them more vulnerable to the financing costs of implementing the tracking and development activities that CEP typically requires (Toxopeus et al. 2021). Consequently, access to capital can boost SME investment in circular economy practices.

Fourth, the findings revealed that access to finance has a direct effect on SMEs’ sustainability performance. Firms with greater access to credit can invest the gathered funds in lowering energy usage, business waste, resource consumption, and the cost of waste treatment, therefore enhancing the organization’s environmental and sustainability performance. This finding is analogous to that of Ullah et al. (2021), who argue that access to internal and external financing leads to superior sustainability performance in SMEs. Prior research suggests that a crucial impediment to the effective implementation of green and sustainable business practices in lean SMEs is a financial constraint, given the absence of immediate quantifiable benefits, high operating costs, and declining sales because of the premium price of green products (Caldera et al. 2019; Kahupi et al. 2021). Fifth, our findings suggest that circular economy practices are significantly associated with firms’ sustainability performance. This finding is in line with that of Corona et al. (2019), Rodríguez-Espíndola et al. (2022), and Walker et al. (2022), who argue that CEP is an important determinant of organizational SP. The negative environmental effect of corporate activities can be minimized by employing circular economy practices (Khan et al. 2022). Circular economy (CE) emerged as a modern ideology that improves businesses’ economic, environmental, and social factors to shift the entire society toward greater sustainability through the engagement of all stakeholders (Dey et al. 2020). CEP is imitable practice that can be transferred across firms. Hence, this finding also validates the PBV, which asserts that imitable and transferrable practices like CEP can drive organizational sustainability performance (Khan et al. 2021c; Tang et al. 2022).

Sixth, we observe a strong positive association between a firm’s absorptive capacity and sustainability performance. This positive association suggests that if a company can effectively manage its capabilities and knowledge resources, it will improve its corporate sustainability. This finding is congruent with numerous studies that have reported a positive linkage between AC and SP over the past decade (Shahzad et al. 2019; Dzhengiz and Niesten 2020; Bhupendra and Sangle 2022). AC contributes to the implementation of sustainable practices since a successful execution involves combining knowledge from many sources, which frequently transcend organizational borders (McWilliams and Siegel 2001; Delmas et al. 2011). Dzhengiz and Niesten (2020) argue that firms with a higher AC can assimilate, convert, and exploit knowledge to generate environmental competencies and capabilities, which drives sustainability performance. This finding supports the DCT since AC is a dynamic capability that facilitates the improvement of managerial competencies and organizational capabilities (Dzhengiz and Niesten 2020) to enhance performance in a changing environment. The finding also demonstrates that CEP mediates the FA-SP and AF-SP associations. This observation resonates with past studies indicating that I4.0 technologies, such as Fintech, are drivers of the circular economy, which, if successfully implemented, may enhance an organization’s sustainability performance (Pizzi et al. 2021; Khan et al. 2022; Tang et al. 2022). Pizzi et al. (2021) conclude that CEBM mediates the association between Fintech adoption and organizational sustainability. Moreover, extant literature suggests that SMEs with better AF can invest in green and CE initiatives, leading to better sustainability performance (Caldera et al. 2019; Jesus et al. 2021). The findings also revealed the moderating effect of AF on the FA-CEP linkage and moderating effect of AC on the CEP-SP association. Since one of the critical features of Fintech is that it can provide access to financial resources to SMEs for CE investments, already having a better AF may dampen the effects of FA on CEP. It could be a reason for the negative moderating effect of AF on the association between Fintech adoption and circular economy practices. The moderating impact of AC supports prior research that suggests that firms with higher absorptive capacity can better implement CEP to ensure sustainability (Dzhengiz and Niesten 2020; Marrucci et al. 2022).

Conclusion and implications

This research aimed to assess the influence of Fintech innovations on the circular economy practices and sustainability performance of Bangladeshi manufacturing SMEs. Furthermore, the moderating impacts of access to financing and companies’ absorptive capacity among the relationships were examined. Our two-staged SEM-ANN investigation demonstrated that Fintech substantially impacts the CEP and SP of businesses. Fintech innovations may mitigate environmental challenges and boost the productivity of businesses with high pollution levels in several ways. Fintech innovation promotes the expansion of green financing via green loans and investments, thus promoting green growth and ecological sustainability. In addition, Fintech contributes to the transition from linear to CE business models by providing SMEs with access to the technologies, such as mobile payment platforms, IoT, and AI, required to achieve the strategic flexibility needed for CE implementation. We observed that enterprises’ access to financing amplifies the influence of FA on SP. In addition, circular economy practices have a favorable effect on the sustainability performance of organizations, and firms’ absorptive capacity strengthens this relationship. Furthermore, better access to financial resources improves CE adoption and manufacturing firms’ ensuing sustainability performance.

Theoretical implications

This study makes several theoretical contributions to the existing literature on Fintech, CEP, AC, and sustainability performance. First, our research enhances the sparse literature on SMEs in emerging economies by extending the conceptual framework of PBV and DCT. Regarding PBV and DCT, this study has significant implications for enhancing CEP and attaining superior sustainability performance in a highly competitive context. Our study extends the research on Fintech adoption and environmental effects (CEP and SP) of firms by merging the two theories and studying the various processes through which access to financing and absorptive capacity affect the outcome. Second, despite the remarkable global growth of Fintech, little is known about the impact of Fintech adoption on a company’s sustainability practices in the age of the fourth industrial revolution (I4.0) (Pizzi et al. 2021). This study offers a better comprehension of how organizations may enhance their sustainability performance by embracing and utilizing financial technology, thereby bridging the research gap. Our research provides empirical evidence to support the claim that deploying financial technology is critical for improving CEP and SP. Our findings suggest that adopting Fintech positively affects CEP and SP, which is compatible with the PBV theory. This conclusion is consistent with results from other regions of the world (Khan et al. 2022; Tang et al. 2022). Thus, it can be argued that the PBV theory is not geographically constrained and may be used worldwide to achieve CEP and SP objectives. Hence, our research expands the scope of the PBV hypothesis.

Third, our research contributes to the extant literature by establishing a direct linkage between CEP and sustainability performance. It implies that an increase in circular economy activities significantly affects the sustainability performance of Bangladeshi SMEs. This finding is a critical contribution to the existing research in developing economies, given the preponderance of studies demonstrating a direct relationship between CEP and SP conducted in developed countries (Dey et al. 2020). Using the PBV paradigm, our finding expands the discourse on the importance of green practices, particularly CEPs, in the sustainability performance of organizations. Fourth, the present work extends the CE literature by assessing the mediating effect of CEP on the interplay between FA and SP. Existing work has predominantly focused on the CEP as a product of I4.0 technologies (Khan et al. 2021c; Tang et al. 2022) and a promoter of organizations’ SP (Dey et al. 2020; Khan et al. 2021b) while ignoring its mediating function. Our study focuses on the mediating impact of CEP between FA and SP. The empirical evidence reveals that CEP positively mediates the relationship between FA and SP. Thus, findings suggest that adopting Fintech enables SMEs to invest in CEP initiatives, including reducing, reusing, and recycling materials which can substantially boost firms’ SP. Given the paucity of studies exploring the crucial mediating function of CEP in the FA-SP relationship, our findings contribute to the literature.

Fifth, this study incorporates AF and AC as crucial moderators and demonstrates that AF directly affects CEP, and AC directly affects firms’ SP. Our research adds to the AF and AC literature by establishing their roles in the transition toward circular economy and sustainability. A voluminous literature has documented financial barriers as one of the significant barriers to a circular economy (Caldera et al. 2019; Jesus et al. 2021). However, in the context of SMEs, this is one of the pioneer studies that empirically examined the role of access to capital in enhancing CEP. Moreover, several studies have established the positive effect of AC on organizational sustainability. We adopted the DC view of firms to explain how AC as a dynamic capability can strengthen the impact of CE on firms’ SP.

Managerial and policy implications

Our study findings have substantial implications for SME managers in emerging economies. We suggest that a combination of Fintech, better access to finance, and absorptive capacity is pivotal for enhanced circular economy practices and sustainability performance of SMEs. Fintech offers a vast array of funding and peer-to-peer lending alternatives, enabling SMEs to alleviate their financial constraints. Managers of SMEs can evaluate whether a company’s activities have a favorable or deleterious effect on each SDG. Using contemporary Fintech innovations like “Robo-advisors,” they can channel the capital of small investors to sustainable projects. Using blockchain technology, SMEs can monitor CE activities and assess the environmental effect of their resources. Fintech cannot, however, improve the sustainability performance of SMEs on its own. To enhance sustainability performance, SMEs must employ green and circular economy practices and develop dynamic absorptive capacity.

Policymakers worldwide have started exploring the enabling role of I4.0 technologies in the sustainable transition of businesses, with a particular emphasis on SMEs, given their pivotal role in a substantial portion of the global economic system. Earlier research advocated the prevalence of obstacles that hinder SMEs’ implementation of environmental practices (Caldera et al. 2019). Our findings indicate the possibilities associated with including Fintech to address such barriers. Therefore, within the framework of policy measures linked with Industry 4.0, Fintech can accelerate the implementation of CEP by SMEs while simultaneously boosting their sustainability performance. Therefore, regulatory bodies in emerging economies must consider integrating technological reforms into their environmental legislation, which will aid in developing crucial mechanisms of incentives and punishments, such as the revocation of company licenses, monetary fines, and carbon tax, in the event of environmental law violations. Lastly, laws and regulations should emphasize the localization of fintech and sustainability goals following the capabilities and ambitions of individual nations. As each nation’s capability to embrace and prepare for fintech and sustainability practices vary, the government must understand these ideas and implement them into national and local economic goals. With guidance and regulations in place, rules may be devised and enforced more flexibly, boosting the SME sector’s possibility of adopting Fintech effectively and complying with the nation’s sustainability objectives.

Limitations and future research avenues

Despite providing potential directions for further research, this research is not without its limitations. First, since we employed a cross-sectional research methodology, we could not confirm the variables’ causal implications. Future researchers might adopt a longitudinal design to determine the time lag and long-term effects. Next, there is a dearth of literature on the association between Fintech and CE. Most of the literature predominantly focused on the association between CE and Industry 4.0 technologies in general or blockchain technology. Thus, we call for future research to explore the effect of Fintech on CE and sustainability practices in different contexts.

Additionally, our work employs variance-based structural equation modeling (PLS-SEM) and an ANN for data analysis and model validation. Covariance-based structural equation modeling (CB-SEM) and contemporary techniques like fuzzy-set qualitative comparative analysis (fsQCA) with a larger sample size could also be used in the future to corroborate the findings. Additionally, this study did not employ robustness checks for the SEM findings. Future researchers may assess unobserved heterogeneity and endogeneity issues. Furthermore, we incorporated SP as a single construct in our conceptual model. However, a growing body of research acknowledges that SP is a multidimensional concept that combines the triple bottom line of environmental, economic, and social performance (Jha and Rangarajan 2020). Consequently, future researchers may employ a multidimensional SP construct to investigate how FA and CEP influence the sustainability performance of firms. In conclusion, despite a growing number of studies on Industry 4.0 technologies emphasizing the importance of examining the circular economy practices and sustainability performance of SMEs from a multi-country viewpoint, utilizing data from multiple sectors, our exploration is limited to a single country and a single industry. Future research can expand our model to different sectors and countries to confirm the theory and findings’ universal applicability and validity.

Author contribution

Abu Bakkar Siddik: conceptualization, investigation, methodology, software, formal analysis, data curation, writing — original draft, and writing — review and editing. Li Yong: supervision, visualization, validation, and writing — review and editing. Md Nafizur Rahman: investigation, data analysis, methodology, software, and writing — review and editing.

Data availability

The data that support the findings of this study are available from the corresponding authors (A.B.S.) upon reasonable request.

Declarations

Institutional review board statement

Not applicable.

Informed consent

Not applicable.

Conflict of interest

The authors declare no competing interests.