2.1 GSCPs and CE

CE, which involves buying, recycling, and reusing resources, has gained more importance due to the COVID-19 epidemic (Donkor et al. 2021). Incorporating GSCP could simplify the transition towards embracing CE. The notion of GSCP emphasizes sustainable management within the organizational framework (Gazi, Al Masud, et al. 2024). Katz-Gerro and López Sintas (2019) identified that firms will develop the capability to reduce resource usage, protect the environment, and move forward with the CE transition if they adopt GSCP. GSCP is designed to gain environmental benefits from CE principles. Several researchers suggested that GSCP is a crucial driver for CE transition and has possible synergy between these two concepts (Abdallah et al. 2024). Manufacturing firms promote GSCP through efficient procurement and resource management (Bressanelli et al. 2018; W. Zhang et al. 2020). Firms require strong governance to integrate green supply chains and sustainability practices.

2.2 GI and CE

To achieve sustainability in the long run, organizations need to invest in projects that are environmentally friendly and socially responsible. Transforming from traditional to circular business models needs GI for infrastructure and technology upgrades (Rahimi et al. 2023). Banks can contribute to CE by sharing information, offering green finance, supporting businesses, educating employees, and minimizing waste (Ozili and Opene 2022). Financial institutions can align GI policies with CE financing. Firms should embrace a CE to benefit from tax exemptions (Ozili 2021). Firms are transformed into CE model require varied sources of financing and GSCP (Saarinen and Aarikka-Stenroos 2023; Agyabeng-Mensah et al. (2021). GI facilitates the implementation of sustainable strategies in the fashion industry (Kolya and Kang 2023). As stakeholders emphasize sustainability, GI appears to be a catalyst for an environmentally friendly and socially equitable future. Investors can promote eco-friendly production processes and ensure long-term profitability by lowering waste, ethical labor practices, and energy-efficient technology (Ikram et al. 2021; Saha et al. 2021).

2.3 GHC and CE

Ability–motivation–opportunity (AMO) theory explains how HRM practices can improve human capital and capabilities by reducing waste, improving productivity and quality, and ensuring higher profit (Shoaib et al. 2021). Green skills and training enhance employees' ability to implement CE practices. Incentives and a sustainability-driven culture foster motivation and encourage proactive engagement (S. H. Ali et al. 2024; Usman et al. 2022). Opportunity is created through supportive policies and leadership, enabling green innovations like recycling and sustainable supply chains. By leveraging GHC across these dimensions, firms can effectively transition to CE, ensuring long-term sustainability. Green human resource management (GHRM) is the summation of organizational commitments, knowledge, wisdom, creativity, and experience in green innovation and environmental protection (Chen 2008). Rubel and Rimi (2024) studied the role of GHRM and the perception of employee CE practices in the RMG sector of Bangladesh. Their study found that GHRM has a direct impact on employees' CE practices perception. Technologically sound human capital possesses environmental knowledge, adapts to new tech, and has strong interpersonal skills for unforeseen contingency (Karmaker et al. 2023). Firms' dynamic capacities in the current revolution promote CE adaption, demanding technologically proficient individuals (O. Khan et al. 2020). Eco-friendly supply chain strategies and technology-oriented human resources enable the RMG firm to manage sustainably and gain a competitive advantage. Green human resources strengthen managerial fundamental competencies in pursuit of CEA (Wu et al. 2021). Researchers found that employees with technological expertise strengthen firms' dynamic capacities for CEA (Kirchherr et al. 2017; Santa-Maria et al. 2022). Yusliza et al. (2020) revealed that only technologically savvy personnel could provide sustainable management. Sustainable RMG practices involve recruiting employees with quick learning abilities and technological adaptability (Bag and Pretorius 2022). RMG's sustainable practices, such as using eco-friendly materials, minimizing waste through efficient production, and recycling fabrics, align well with CE's goals. For instance, some garment manufacturers are establishing closed-loop systems that collect, disassemble, and reuse old garments to create new ones (Shafique et al. 2023; Ozili and Opene 2022; Rahimi et al. 2023). This aligns with the CE principle of a product life cycle that does not end with disposal but instead transitions back into production. In addition, by adopting eco-design principles and using sustainable raw materials like recycled fibers, RMG companies help reduce resource consumption and waste while promoting environmental and economic sustainability (Sarkar et al. 2020a, 2020b).

2.4 SRMG and CE

SRMG sector entail producing goods without harming the environment (Yadav et al. 2020). Reducing, reusing, and recycling could aid firms in adapting to the CE, managing resource and environmental challenges, and promoting sustainable growth (Gonçalves et al. 2022; Geissdoerfer et al. 2017). SRMG sector can emerge as an effective strategy to leverage CE capabilities, addressing global concerns about resource and material degradation (Dagilienė et al. 2023; Liu et al. 2019; Yadav et al. 2020; Jinru et al. 2022).

Several studies indicated that GSCP, GI, and GHC influence CEA (Abdallah et al. 2024; S. H. Ali et al. 2024; Yadav et al. 2020; Jinru et al. 2022). Organizational managers should develop circular practice action plans to optimize resources for sustainable management (S. Khan and Haleem 2021). Firms can hire individuals with strong technological skills and environmental awareness to promote sustainable practices (Holgersson 2013). This study postulates that the influence of these factors on CE adaptation is not directly related rather it is indirect. By embracing sustainability practices through GHC, GI, and GSP, the RMG industry can directly and indirectly move towards the transition to ce (Assmann et al. 2023; Dagilienė et al. 2023; Charitou et al. 2021). Thus, the following conceptual framework (Figure 1) is developed based on the above literature studies.

3.1 Sampling Design and Procedure

An opinion survey was conducted on the personnel (mostly the mid-level/higher level employees of the RMG sector) with relevant knowledge of sustainability practice and CE. This study considers the firms that are registered with the Bangladesh Garments Manufacturers and Exporters Association (BGMEA). The sample firms were determined based on LEED green garment factories certified by the U.S. Green Building Council (BGMEA 2021). These sample RMG firms were selected because it is assumed that their employees and management will know about sustainability, sustainability reporting, and circular business models. The questionnaire was devised after consultation with the academic researchers with careful consideration of layout, word selection, readability, and construction of scale items. The questionnaire underwent back translation to ensure a clear understanding.

3.2 Data Collection

A pilot survey was conducted to validate the questionnaire. After accommodating the feedback from the pilot survey, the questionnaire was distributed among the potential respondents. A total of 650 invitations were sent by email and face-to-face surveys based on convenience and snowball nonprobability sampling techniques. Convenience sampling allowed for quick and cost-effective data collection, ensuring that a substantial number of responses could be gathered within a limited timeframe (Sherman 2024). In addition, in cases where access to a well-defined population is challenging, snowball sampling is a practical approach (Gazi, Al Masud, et al. 2024). This technique enabled the identification of additional respondents through referrals from initial participants, which was particularly useful in reaching individuals who might not have been accessible through direct invitations. There were varied opinions found on required the sample size for PLS-SEM analysis. According to Gorsuch (1983), there should be at least five observations for each construct. Harris and Schaubroeck (1990) and Kline (2023) suggested that the size of the sample is supposed to be not less than 200 to ascertain robust SEM. A total of 613 responses were collected between September 2023 and February 2024. Finally, 589 responses were retained, and the rest were discarded for incompleteness.

3.3 Overview of Analysis

For analytical purposes, PLS-SEM is used. Additionally, NCA is used to identify the necessary condition of any output under study. PLS-SEM is employed in this study to analyze the relationships between predictor and outcome variables. PLS-SEM is a variance-based structural equation modeling technique widely used in social sciences and business research due to its ability to handle complex models with multiple constructs and indicators (Hair et al. 2019). It is particularly beneficial for maximizing predictive accuracy, handling small to medium sample sizes and analyzing nonnormal data distributions (Haji-Othman et al. 2024; Richter and Tudoran 2024). Previous research, such as studies by Armutcu et al. (2024), Mensah et al. (2024), and Castellano et al. (2022), has extensively applied PLS-SEM. In addition, NCA is employed alongside PLS-SEM to identify critical conditions required for achieving specific outcomes. NCA determines whether certain conditions are essential but not necessarily sufficient for a particular outcome to occur (Dul 2016). It identifies bottlenecks—conditions that must be met before an expected outcome can be realized. This approach is particularly useful in fields where performance thresholds must be achieved (Escadas et al. 2023; Dul et al. 2023; Sukhov et al. 2023; Vaithilingam et al. 2024; Hauff et al. 2019).

3.4 Instrument Development and Measurement

The items of the measurement model were derived from existing literature, emphasizing the incorporation of established research concepts and indicators. We used a 7-point Likert scale to measure the exogenous factors transparently within the study and to gather the varied stimuli of the opinion of the participants (Gazi et al. 2025; Sarstedt et al. 2023). The constructs used in this study are GSCP, GI, GHC, SRMG, and CEA. The items for GSCP were derived from previous study by Islam et al. (2024). The metrics for GHC were adopted from Shoaib et al. (2021) and Marrucci et al. (2021). The elements of GI were sourced from S. Ali et al. (2022) and Rahimi et al. (2023). Items for SRMG and CEA were adopted from (Shafique et al. 2023; Ozili and Opene 2022; Rahimi et al. 2023; Marrucci et al. 2021; Wu et al. 2021; Sarkar et al. 2020a, 2020b). The study scrutinized data adequacy for factor analysis through correlation matrix measurement, revealing correlations and determining the number of retained factors via eigenvalue parameters (Cattell 1966).

3.5 Findings and Analysis

3.6 Demographic Information

Our dataset consists of 589 respondents from diverse managerial positions and sectors of the Bangladeshi RMG industry. The majority of the workforce (60.45%) are male employees. More than half of the managers (56.54%) have attained a graduate degree. A large proportion (37.52%) of the study participants reported having 3–5 years of managerial experience in their current or previous role. The details about the demographic profiles are given in Table 1.

3.7 Convergent Validity

Table 2 depicts acceptable convergent validity, with Cronbach's alpha values (0.800–0.890) meeting the criterion composite reliability (CR) > 0.7 (Cortina 1993). In addition, factor loading standardizations for observed variables go beyond the recommended value (≥ 0.6) as posited by (Bagozzi and Yi 1988). Based on Fornell and Larcker (1981), all items' factor loading must be greater than 0.60 to establish convergent validity. The reported CR values exceed the recommended criterion, such as CR > 0.7, as suggested by Hair et al. (2011). Additionally, the AVEs exceed the threshold value of 0.5, as proposed by Hair et al. (2011), indicating satisfactory convergence in our study.

3.8 Discriminant Validity

Following the recommended approaches, Hair et al. (2017) this study rigorously assessed discriminant validity using Fornell and Lacker (FL) criteria and the heterotrait–monotrait (HTMT) ratio.

Table 3 depicts the findings of the FL criterion analysis, indicating that discriminant validity is confirmed as the square roots of AVE exceed off-diagonal values.

By the criteria by Hair et al. (2017) and Henseler et al. (2015), Table 4 displays an HTMT correlation ratio below the specified threshold of 0.85, meeting the required standard (Gazi, Al Masud, et al. 2024). Hence, this confirms discriminant validity and ensures the distinctiveness of constructs in the research context.

3.9 Structural Model and Results of Hypothesis Testing

After validating the outer model's reliability and addressing no multicollinearity issues, this study evaluated proposed relationships among inner models. Table 5 presents the results of the direct hypothesis analysis of the β values, t values, and p values. In this study, the significance of each structural path (Figure 2) was typically determined through the bootstrapping procedure.

Table 5 displayed that GSCP significantly influenced CEA (β = 0.26, t statistics 3.653, p < 0.001) and SRMG (β = 0.15, t statistics 6.59, p < 0.001). Thus, the findings supported H1 and H2. GI has a significant positive influence on CEA and SRMG (β = 0.10, t statistics 2.489, p < 0.000) and SRMG (β = 0.28, t statistics 7.037, p < 0.001). Hence, H3 and H4 are supported. Moreover, TEHC positively influences CEA (β = 0.19, t statistics 4.673, p < 0.000) and SRMG (β = 0.33, t statistics 8.441, p < 0.000). Thus, H5 and H6 were supported. In conclusion, SRMG significantly influences CEA (β = 0.35, t statistics 8.733, p < 0.000). Therefore, there is empirical support for H7.

This study also assessed the indirect effects of GSCP, GI, and GHC on CEA through SRMG. Table 6 outlines the findings of mediation employing the product coefficient method. Findings indicate significant indirect effects in all paths, providing robust support for mediation, with statistical significance at p < 0.05 (Zhao et al. 2010; Preacher and Hayes 2004). Moreover, we employ the variance accounted for (VAF) approach to assess the presence of partial, full, and no mediation within the model. VAF less than 20% implies no mediation, more than 80% suggests full mediation, and 20%–80% indicates partial mediation (Hair et al. 2017).

3.10 NCA

To identify the necessary conditions, the guidelines provided by Dul (2016) were employed based on three primary criteria. Firstly, the expected relationship between predictor and outcome variables needed a sound theoretical base. To ensure relevance, it is required to have a positive effect size. Thirdly, to guard against false positives and Type 1 errors, it subjected the conditions to rigorous testing against the null hypothesis. This involved utilizing a bootstrapping approach and a permutation test (Dul et al. 2023; Rahman et al. 2024). In this study, the latent variable scores from PLS-SEM are employed for NCA. The ceiling regression-free disposal hull (CR-FDH) line is suitable for handling discrete data with multiple levels (Dul 2016; Basco et al. 2022). This study also looked at the effect sizes of latent variable scores (with d values between 0 and 0.1) and checked for statistical significance using a permutation test with 10,000 random samples, which is similar to how Dul et al. (2023) did it. The results indicate (Table 7) that several conditions were necessary for the RMG to adopt a circular business in Bangladesh. In certain cases, the effect sizes exceeded zero. Among these conditions, CEA with GSCP demonstrated the largest effect size (d = 0.98, p < 0.01). CEA with GI (d = 0.75, p < 0.05), CEA with GHC (d = 0.067, p < 0.01), and SRMG (d = 0.045, p < 0.05) were also identified as necessary conditions. However, due to the small effect sizes of GHC and SRMG, these conditions were considered less important. The virtual inspection of the plot in Figure 3 also pointed out that.

To identify the essential role of GSCP, GI, GHC, and SRMG in grasping the benefits of CEA, a bottleneck table (Table 8) is constructed. This table outlines the minimum values (%) required for GSCP, GI, GHC, and SRMG to attain CEA (80% and above). The minimum levels are 1.18% for GI, 1.35% for SRMG, 6.62% for GHC, and 0.34% for GSCP. The bottleneck table clearly demonstrates that achieving higher levels of CEA requires progressively greater contributions from GI, SRMG, GHC, and GSCP. Each factor's contribution becomes increasingly critical as the desired level of CEA rises, with significant jumps, particularly evident at the 90% and 100% marks (Figure 3). These findings underscore the pivotal roles these green practices play in advancing towards a CE in the RMG sector.

5.1 Theoretical Contribution

This study makes significant theoretical contributions to our understanding of sustainable practices in the RMG industry, particularly in the context of emerging economies like Bangladesh. Firstly, it enhances our understanding of the effective integration of CE principles into the RMG industry in emerging economies. The existing literature often focuses on CE models within developed nations, leaving a gap in practical application and strategies suited for developing countries like Bangladesh. By exploring the unique challenges and opportunities within the Bangladeshi RMG sector, this study provides a framework for other emerging economies, thereby broadening the theoretical discourse on global sustainability practices. Secondly, the study delves into the direct impact of GSCP, GI, and GHC on SRMG and CEA. While previous research has frequently examined these variables in isolation or in different contexts, this study synthesizes them within the Bangladeshi RMG industry's specific environment. It enriches the theoretical understanding of how these green practices interact and contribute to sustainability and circularity, providing a nuanced view that can inform both academic research and industry practices. Lastly, the study addresses a critical gap in the literature by investigating the mediating role of SRMG concerning GSCP, GI, and GHC. This mediation analysis offers deeper insights into the mechanisms through which green practices influence the adoption of CE principles. The findings will aid theorists in comprehending the crucial role of sustainable practices in the RMG sector and their potential to improve wider environmental and economic results. This contribution is particularly valuable for developing comprehensive sustainability models that incorporate multiple interrelated factors.

5.2 Managerial Implications

The findings of our study elucidate a practical implication for organizational management, corporate governance, and policymakers in Bangladesh. Our empirical study establishes that transformation towards a CE needs to integrate diverse GSCP strategies, GI, and the strategic utilization of GHC. This transformative approach is necessary for keeping sustainable RMG practices within the garments industry. CE adaptation can mitigate resource shortages and create a competitive advantage for firms. Furthermore, CEA is essential for firms' sustainability, especially in South Asia's vulnerable ecological context, where stakeholders demand conventional firms to adopt green practices for sustainability and circularity. As an emerging South Asian economy, Bangladesh depends heavily on its RMG sector, which our findings show has a considerable influence on how to adopt circular transition. We propose a framework for garment industry transformation that needs to apply GSCPs throughout the distribution process, government and financial institutions should come forward with more GI options and tax benefits. Lastly, organizations should build awareness and train their human capital. It also suggests policy interventions for the government to support the adoption of sustainable RMG management practices.

5.3 Limitations and Future Research Guidelines

The study provides significant contributions to business and society while also recognizing its limitations and offering suggestions for future research. This study's findings, based on the garment industry only, may not generalize to all manufacturing industry contexts. Future research should collect data from diverse sectors and regions to inform universal decisions for emerging economies in South Asia. Moreover, the CE is a novel concept in Bangladesh. Firms are still in a very nascent stage of adopting it. But the performance of each construct has not been measured yet. To find out how far the CE has really come, performance matrix analysis should be used in future research. Furthermore, future research could broaden its scope by conducting multigroup analyses across diverse regions. Future research should use mixed methods to verify the validity of the findings and examine the interaction of GSCP, GI, and GHC with corporate social responsibility.