Investigating the Implementation of Circular Economy in the Textile Industry in Indonesia

2.1 Business analytics capability and circular economy implementation

In the digital era, it turns out that the issue of sustainability and CE has become a strategic issue and the main goal of the company. The various resources owned by the company must be utilized as effectively as possible so that the decisions made are correct and can benefit the company. To realize this, companies need to have BAC. BAC is a collection of technologies, methods, and applications that enable business data analysis to promote better and data-driven decisions [10, 11]. BAC can accelerate the CE transition [6] by helping companies expand their decision-making scope, providing insights and options that were previously unavailable [34], increasing the speed of response time, effectiveness, and efficiency in adapting environmental changes [21].

With BAC, companies can effectively mobilize, deploy, and utilize resources and align planning with their business strategy to improve company performance [35, 36]. The challenge for companies is that they need to identify organizational resources that can support transformation when using BAC. Companies need to develop a combination of tangible, intangible, and human resources [6, 22-24]. BAC can also enhance a company's ability to operationalize its circular strategy and overall CEI [6, 37]. The role of BAC can accelerate a company's CE adoption and realize the company's business value. According to Kristoffersen et al. [6], BAC strengthens the implementation of circular strategies and firm performance, particularly in terms of competitiveness, corporate reputation, financial performance, and EP.

H1: BAC has a significant positive influence on CEI.

2.2 Green human resource management and circular economy implementation

The literature on CE is growing rapidly, but the implementation of CE in companies rarely intersects with GHRM. The understanding of CE principles is very limited in the wider community [15], whereas the development of CE competencies is needed [14]. Therefore, Jabbour et al. [18] developed a theoretical framework that conceptualizes GHRM for CE. GHRM and CE have certain similarities and produce certain synergies that need to be explored so that these synergies can be enhanced [18]. GHRM practices can improve corporate sustainability, especially in terms of better BP, EP, and SP [38]. GHRM should be considered as part of the concept of sustainability and CE business models that will affect company performance [16, 17].

However, the average company ignores the key role of GHRM in the organization as one of the determining factors in the CE model. Therefore, Jabbour et al. [18] proposed an integrative framework to link GHRM and CE at the organizational level. HRM enables CE by aligning GHRM practices and dimensions of recruitment and selection, training, performance appraisal, rewards, culture, teamwork, and empowerment with the CE business model. Jabbour et al. [18] argued that the implementation of workplace-based GHRM practices can positively influence the development of CE strategies in organizations. Dimensions that focus on such things as organizational culture, empowerment, and teamwork enhance the success of sustainability initiatives. It can be concluded that GHRM can aid the implementation of CE business models [38-40]. In addition, top management commitment, managerial leadership, and employee motivation can facilitate GHRM in positively influencing CEI and firm performance [40]. GHRM policies can facilitate transformation in creating an environmentally friendly organizational culture [5]. This turns out to have an impact on the application of GHRM practices on the implementation of CE principles. Jabbour et al. [18] argued that CE and the adoption of CE business models can be achieved if workplace-based GHRM practices are used and integrated.

H2: GHRM has a significant positive relationship with CEI.

2.3 Digital supply chain towards circular economy implementation

CE is a topic that is widely discussed in various literature amid uncertain climate conditions. Companies are required to take responsibility for the safety of the earth, and CE is the right solution to overcoming these problems [41, 42]. CE is a system to minimize the waste of resources and reduce waste disposal by recycling [13, 30, 31, 43]. There are several efforts to successfully implement CE in a company, one of which is digitalization [12, 13, 44]. In some cases, CEI in a company requires changing the business model, namely by digitizing the business. In China, digitalization has helped and facilitated cooperation between the private sector and the government in CEI [30]. Digitalization is an effective solution for optimizing the implementation of a CE.

Several technologies have been used to carry out waste management, such as the “Pay as You Throw” system, the “Know as You Throw” scheme, and the use of 'Radio Frequency Identification' (RFID), which supervises and monitors waste activities in the industry [31, 45]. Broadly speaking, digitization brings one major change, namely efficiency. Digitalization utilizes information and communication technology to add value to a product or company system [12, 44]. Without digitalization, the implementation of CE will be slow, and the impact will be small [46]. Digitalization is also very suitable for post-pandemic conditions so that post-pandemic economic problems can be resolved [31]. Agarwal et al. [47, 48] argued that what can optimize the implementation of CE is digitalization. Therefore, companies must be able to identify and optimally utilize their supply chain management. This is important because can be used to improve CEI in all value chain activities from upstream to downstream. When the use of digitalization is well integrated and supported by quality human capital, DSC will be able to improve BS in the long term [32, 33, 49-51]. The adoption of DSC is expected to strengthen the integration and agility of the supply chain so that customers and stakeholders will be satisfied and can make repeat transactions with the company [9, 52, 53].

H3: DSC has a positive effect on CEI.

2.4 Circular economy implementation on business sustainability

Companies today are given the responsibility to maintain economic prosperity and improve environmental quality. One of the efforts that can be made is to design and implement CE strategies in every business activity [3, 5, 7, 54]. In reducing resource wastage and emissions as well as waste disposal, CEI can be used to accelerate it [55, 56]. The implementation of CE has a principle known as the 4R framework, namely reduce, reuse, recycle, and recover [7]. When companies can implement CE properly and correctly, business performance (BP) will improve [4, 6, 57, 58]. Companies can operate efficiently by carrying out several activities, such as reusing raw materials, carrying out recycling processes, and recovering resources used in business processes from upstream to downstream [59].

Shahzad et al. [60] and Yu et al. [61] stated that the implementation of CE can improve business efficiency in all aspects such as production efficiency and environmental cost efficiency. The recycling process in CE can save scarce resources through the reuse of materials used in the production process. On the other hand, recycling also saves land that is usually used for industrial waste disposal, thus eliminating harmful gases such as methane that arise from waste [62-68].

Various literature has also introduced the concept and implementation of CE in producing environmental sustainability performance [3, 4, 56, 69] to social sustainability [70-72]. Several studies agree that the implementation of CE can improve EP [4, 37, 56, 63, 69, 73]. Schwarz et al. [69] and Vollmer et al. [74] argued that the recycling process in CEI can have a major impact on EP. CE can reduce the adverse impact of business operations on the environment. The implementation of CE minimizes producers' dependence on resources through recycling and remanufacturing. This practice provides many benefits for companies in the form of performance efficiency and improved EP [18, 63].

Business, environment, and society are three aspects that must be considered by companies. Business is related to profits and company performance; the environment is related to government regulations and also the company's obligation to participate in preserving the environment. No less important is the social aspect, which is the key to synergizing with other parties and a form of social responsibility [72, 75]. Padilla-Rivera et al. [70] collected 31 social aspects that can be developed through CE. Of these 31 aspects, the two most affected aspects are society and practice and decent work. More specifically, the most discussed aspect is about employment. The social dimension is important due to the need for cooperation and interdependent relationships between various stakeholders [64, 65]. Saidani et al. [67] and Moreau et al. [68] recommended integrating social dimensions in the implementation of CE to create good SP and EP.

Meanwhile, Jabbour et al. [18] explored the human side of the CE by connecting the CE and GHRM. Several studies prove that CE can have a positive impact on SP [37, 70, 71]. There are several social benefits obtained in the implementation of the CE, namely creating new employment opportunities through new uses of the value embedded in resources, increasing the sense of community, cooperation, and participation through the concept of the sharing economy, and product functions and services are more likely to be in the public interest than individuals [71]. Tetteh et al. [37] in their findings, confirmed that there is a significant partial mediating role of CE on the influence of BAC and BAC sustainable performance.

H4: CEI has a positive effect on BP.

H5: CEI has a positive effect on EP.

H6: CEI has a positive effect on SP.

H7: CEI mediates BAC on BP.

H8: CEI mediates BAC on EP.

H9: CEI mediates BAC on SP.

H10: CEI mediates DSC on BP.

H11: CEI mediates DSC on EP.

H12: CEI mediates DSC to SP.

H13: CEI mediates GHRM on BP.

H14: CEI mediates GHRM on EP.

H15: CEI mediates GHRM to SP.

Corsini et al. [76] recommended that since the CE is systemic, it should involve a holistic approach, one of which can be analyzed by clustering the characteristics of consumers or producers. The results of their research even prove that there is a significant gap between circular demand and supply. In shaping circular production and consumption, linear producers make up the largest group of companies, while linear consumers represent the smallest segment. Her research underscores the need for comprehensive involvement of both producers and consumers. The research also proposes to analyze the differences in the relationship between variables (covering all the variables studied, i.e., BAC, GHRM, and DSC) between the subgroups studied (both fashion and clothing materials) in CEI. It is also important to look deeper into CEI when it is associated with contingency variables such as BAC, GHRM, and DC.

H16: There is a significant difference in the relationship between variables (covering all variables studied, i.e., BAC, GHRM, and DSC) between the subgroups (both fashion and clothing material) in CEI.

4.1 Respondent profile

The profile of respondents in this study reflects the diverse characteristics of textile MSMEs in West Java and Central Java (fashion and clothing materials) (Table 1). The majority of respondents were female (56.6%), indicating the significant role of women in the textile industry. By age, the oldest group (49-66 years) dominates with 44.6%, indicating that MSME players in this sector are still dominated by the older generation. This is consistent with the distribution of birth years, where the majority were born between 1943 and 1981 (73.7%).

Table 1. Respondent profile

In terms of position, there is almost a balance between owners (36.7%) and managers (36.3%), but interestingly, 27.1% are both owners and managers, reflecting the simple business organization structure typical of MSMEs. In terms of education, the majority of respondents have a secondary education background (high school, 62.6%), while only 37.4% have tertiary education, which may affect the level of adoption of digitalization and strategic managerial practices such as CE.

The types of MSMEs are almost equally divided between fashion products (51%) and clothing materials (49%), allowing for proportional comparison in the multigroup analysis (MGA). Most MSMEs have a small number of employees (2 to 6 people, 67.7%), as well as varying annual turnover, with the largest group being in the lowest category (IDR 10 to 107.6 million per year). This data suggests that the textile MSMEs in the study sample are generally small-scale, directly managed by their owners, and in a demographic context that has the potential to be empowered through strategies based on digitalization, green management, and CE.

4.2 Convergent validity and reliability test

The results of validity and reliability testing (Table 2) show that all constructs in the research model have met the criteria for excellent measurement quality, both in the overall sample and in each group of fashion and clothing material MSMEs. All outer loading (OL) values are above the 0.7 threshold, which indicates that each indicator has strong indicator validity for its construct. The Average Variance Extracted (AVE) of all constructs is also well above the minimum limit of 0.5, which indicates convergent validity has been achieved, meaning that more than 50% of the indicator variance is explained by their respective latent constructs. In addition, the very high Composite Reliability (CR) and Cronbach's Alpha (CA) values (the majority close to or above 0.9) indicate that each construct has excellent internal reliability and strong inter-item consistency. Overall, these results ensure that all constructs used in the model are feasible and reliable for further analysis in structural testing and MGA.

Table 2. Convergent validity and reliability test

4.3 Discriminant validity

The results of the discriminant validity test through two approaches—the Fornell-Larcker Criterion and Heterotrait-Monotrait Ratio (HTMT)—show that all constructs in the model have met the discriminant validity requirements, both in the complete model and the fashion and clothing materials subgroup (Table 3). In the Fornell-Larcker criterion, each AVE square root value (shown on the diagonal of the table) is higher than the correlation between other constructs in each row and column, indicating that each construct is sufficiently unique and does not overlap conceptually. Meanwhile, the HTMT values across constructs remain below the conservative threshold of 0.85, confirming that there is no latent multicollinearity problem between constructs. This indicates that variables such as BAC, GHRM, CEI, and performance outcomes (BP, EP, SP) are truly distinct from each other empirically. In particular, these results provide a strong foundation for further structural analysis, including inter-construct relationship testing and MGA in the next phase.

Table 3. Fornell-Larcker discriminant validity test & HTMT ratio

Note: BAC = Business Analytics Capability, BP = Business Performance, CEI = Circular Economy Implementation, DSC = Digital Supply Chain, EP = Environmental Performance, GHRM = Green Human Resource Management, SP = Social Performance

4.4 Model fit test

The results of the model fit assessment using the Standardized Root Mean Square Residual (SRMR) indicate that the research model demonstrates a very good level of fit (Table 4). The SRMR values for both the saturated model (0.041) and the estimated model (0.048) are well below the recommended threshold (≤ 0.08). This suggests that the discrepancy between the observed and model-implied covariance matrices is relatively small, indicating that the model adequately represents the empirical data. Therefore, the structural model can be considered to have satisfactory goodness of fit and is appropriate for further analysis of the relationships among the constructs.

Table 4. Model fit test SRMR

4.5 Hypothesis model

The results of hypothesis testing show strong empirical support for all direct and mediation relationships proposed in the research model so that all hypotheses are accepted (H1 to H15 are accepted) (Table 5). In general, all tested paths have significant t values (t > 1.96) and positive β (path coefficient), indicating a meaningful and unidirectional influence between variables. This indicates that BAC, GHRM, and DSC play an important role in driving CEI, which in turn has a positive impact on various variables in BS: BP, EP, and SP. This can be seen more clearly in Figure 1.

Figure 1. Inner model (bootstrapping t-statistics)

Table 5. Structural model path coefficient

Specifically, in the direct hypothesis, all three predictors of CEI showed significant influence, with the highest contribution coming from GHRM (β = 0.414; t = 6.057), followed by BAC (β = 0.396; t = 5.862) and DSC (β = 0.310; t = 4.287). This indicates that environment-based and environmentally sound managerial approaches are more dominant in driving CEI. When examined at the group level, the contribution of DSC to CEI was greater in the fashion segment (β = 0.423) compared to clothing materials (β = 0.310), suggesting a variation in the role of DSC depending on the type of industry analyzed.

Meanwhile, the influence of CEI on the three performance variables also proved significant. CEI contributes the most to BP in the full sample (β = 0.476), but the highest contribution generally occurs to EP, especially in the fashion (β = 0.455) and apparel (β = 0.453) segments. This suggests that the implementation of a CE has the most tangible impact on improving environmental sustainability. However, the impact on SP also remains positive and significant, albeit slightly lower (β = 0.277 in the full sample). On the mediation path, it is found that CEI acts as a significant mediator in all indirect relationships. The strongest mediation is reflected in the paths GHRM → CEI → SP (β = 0.532; t = 7.352) and DSC → CEI → SP (β = 0.431; t = 6.846), indicating that HRM policies and digitalization indirectly strengthen the role of circular practices. The differences between subgroups are also interesting to observe. For example, in the apparel segment, the paths BAC → CEI → SP (β = 0.537) and GHRM → CEI → EP (β = 0.424) show stronger effects than in the fashion segment. This signals that the apparel sector is more sensitive to analytics and human capital initiatives in the context of sustainability.

Overall, these results suggest that CEI plays a central role as the main mediating mechanism in bridging various variables (BAC, GHRM, DSC) toward achieving triple-bottom-line performance (economic, environmental, and social). Thus, organizations that want to strengthen sustainability need to strategically integrate BAC, GHRM, and DSC into their CE practices. These findings support the importance of circular-oriented transformation as a cross-functional value-added strategy.

4.6 Multigroup analysis

The results of the MICOM (measurement invariance of composite models) test for compositional invariance showed that the correlations between the variables in the model were relatively stable across the different groups (Table 6). All p-values for the permutation test are above the significance limit of 0.05 (p > 0.05), which means that there are no significant differences in the relationships between variables between the subgroups analyzed (fashion and clothing materials). This indicates that the composition of relationships between variables can be considered invariant or unchanged between groups, which supports using the same model for both groups without worrying about deep structural differences. In other words, despite differences in industry characteristics, the mechanism of relationships between variables remains consistent, allowing for valid comparisons between the groups under study.

Table 6. Compositional invariance

Note: BAC = Business Analytics Capability, BP = Business Performance, CEI = Circular Economy Implementation, DSC = Digital Supply Chain, EP = Environmental Performance, GHRM = Green Human Resource Management, SP = Social Performance

The results of the composite equality of mean values and variance test in Table 7 show that there is no significant difference between the Fashion and Clothing Materials groups, both in terms of means and variances. For most variables, the p-value is greater than 0.05, which means that the difference in means between the two groups is not statistically significant (e.g., BAC with a p-value of 0.456 and CE with a p-value of 0.336). The same applies to the variance, which shows similarity in variability between the groups. Overall, this indicates that the model structure can be applied well to both groups without the need for further adjustments regarding differences in means or variances, thus strengthening the validity of the tests and comparisons between the two groups.

Table 7. Composites equality of mean values and variance

Note: BAC = Business Analytics Capability, BP = Business Performance, CEI = Circular Economy Implementation, DSC = Digital Supply Chain, EP = Environmental Performance, GHRM = Green Human Resource Management, SP = Social Performance

In the MGA test comparing the differences in pathways between the Clothing Materials and Fashion sectors (Table 8), most of the pathways did not show significant differences. For example, the path BAC → CEI showed no significant difference between the two sectors, with a very high p-value (0.762). This confirms that the influence of BAC on CEI is consistent across both sectors. Other pathways, such as CEI → BP and CEI → EP, also showed no significant differences after further analysis, with higher p-values in the new tests (0.214 and 0.310, respectively). Overall, despite the difference in coefficients between the two sectors, the analysis results show that the relationships between the variables are relatively similar, with no significant differences indicating strong sectoral differences in influence. This explanation also rejects hypothesis 16.

Table 8. Multigroup analysis

Note: BAC = Business Analytics Capability, BP = Business Performance, CEI = Circular Economy Implementation, DSC = Digital Supply Chain, EP = Environmental Performance, GHRM = Green Human Resource Management, SP = Social Performance

CE practices carried out relatively still use conventional practices and are not comprehensive. This is because the MSMEs studied have limited resources, including human resources. In HRM practices, what is important is that MSMEs focus on a touch of green management (green orientation) in every practice, such as training, performance appraisal, and compensation based on GHRM practices. Another limitation is in DSC practices, where the most important thing is only emphasizing the use of DSC in certain aspects, such as the adoption of digital product sales, partial implementation in digital operations, and using digitalization in aspects of service to customers and suppliers.

Therefore, the emphasis on CEI is expected to help textile MSMEs to play a role in mediating business capabilities, GHRM practices, and DSC in improving sustainable business. In the practice of CE, it also turns out that MSMEs still apply the most important simple principles, namely reducing synthetic raw materials used in production, distribution, and marketing as much as possible; utilizing leftover materials; recycling existing materials; and engineering in redesigning the use of the remaining existing raw materials.

In the future, research should use mixed methods because researchers will be able to get a more complete and in-depth understanding of the empirical phenomena at hand and can find better solutions. The longitudinal approach is highly recommended because researchers can have a more accurate picture and results of the research theme and can determine and analyze the causal relationship between the antecedents and consequences of CEI. This is because researchers can make repeated and continuous measurements over a long period of time and even years. Therefore, it is also necessary to consider cost and time and to manage the research effectively. The research area only covers two provinces, namely Central Java and West Java. Therefore, the results of the study cannot be generalized because the variation in resources owned by textile MSMEs in the two regions is different.