The effects of sustainability innovation and supply chain resilience on sustainability performance: Evidence from China’s cold chain logistics industry

Abstract

The purpose of this paper is to investigate the impact of sustainability innovation and supply chain resilience on sustainability performance and to predict the interrelationship between economic, environmental, and social performances. Using a cross-sectional quantitative study, an online survey was conducted among the Chinese cold chain logistics companies, and 204 valid responses were collected. Data analysis was conducted using the partial least squares structural equation modelling approach. The results indicated that sustainability innovation and supply chain resilience positively affect three pillars of sustainability performance. Within sustainability performance, economic performance has a significant positive correlation with environmental and social performances, while environmental performance has no effect on social performance. The findings made constructive contributions. Theoretically, this study contributes to dynamic capability theory by predicting the direct impact of sustainability innovation and supply chain resilience on sustainability performance in a model that is relatively new in the literature. In addition, this study examines the relationship between economic, environmental, and social performances in a unique context (i.e. the cold chain logistics industry). Practically, this study will guide practitioners in developing innovative and resilient strategies committed to business sustainability and urge policymakers to develop policies such as subsidies, regulation, and training programmes to promote innovation and collaboration in this industry.

Keywords:

• Cold chain logistics
• sustainability performance
• sustainability innovation
• supply chain resilience
• dynamic capability

REVIEWING EDITOR:

• Virginia Barba-Sanchez, Universidad de Castilla-La Mancha - Campus de albacet, Spain

SUBJECTS:

• Business, management and accounting
• Industry & industrial studies
• Asian studies
• Statistics for social sciences
• Operations management

1. Introduction

It is estimated that about 40% of all food requires a cold chain (Bremer, 2018). However, due to the lack of Cold Chain Logistics (CCL), around 526 million tonnes of food loss are generated annually (Sarr et al., 2021). In the CCL, refrigerated facilities and equipment consume large amounts of electricity and energy, generating considerable environmental problems (Kumar et al., 2022a, 2022b). Also, the lagging development of CCL can cause social issues like food security and vaccine safety (Kumar et al., 2022b). Thus, sustainability is one of the key issues facing the current CCL industry (Arora et al., 2023). As the world’s environmental authorities begin to assess the impact of cold chain practices, more and more Cold Chain Logistics Companies (CCLCs) are actively taking social and environmental responsibility to achieve sustainable development based on the Triple Bottom Line (Liao et al., 2023).

Facing pressure from stakeholders, CCLCs in emerging countries urgently need to explore supply chain practices that are committed to achieving sustainability (Eikelenboom & de Jong, 2019; Shang et al., 2020). In this regard, Afeltra et al. (2023), through an evaluation of 1,108 articles, argued that green innovation is one of the key factors in achieving environmental and economic success, but it does not simultaneously address social sustainability as sustainability innovation does. Farooq et al. (2024) also observed that sustainability innovation can not only improve the environmental sustainability of non-financial sector firms in BRICS countries but also increase the financial efficiency of their economies. Su et al. (2023), who did a survey of experts in China’s agricultural products cold chain, found that sustainability innovation is a key success factor in achieving the sustainable development of agricultural products cold chain, and tended to believe that firms that implemented this strategy can better meet the demands of major stakeholders. As an active and holistic innovation, sustainability innovation enables firms to adapt to changing environments and reduces negative environmental and social impacts by responding to stakeholders’ needs (Wong et al., 2022). It is worth noting that in addition to emphasising the role of sustainability innovation, Taghikhah et al. (2023) also considered that the current food cold chain design requires a shift from ‘time-cost-quality’ to ‘resilience-sustainably-trust’. As a dynamic adaptability, supply chain resilience facilitates firms to prepare, withstand risky events, and return to their original or better state (Chowdhury & Quaddus, 2017; Ponomarov & Holcomb, 2009; Wong et al., 2022). Khan and Ali (2023), He and Liu (2023), and Singh et al. (2023) also claimed that supply chain resilience can effectively resist the interference of disruptive events such as public security emergencies and ensure the sustainable operation of CCL.

Although sustainability innovation and supply chain resilience strategies contribute to achieving CCL’s sustainability performance, most studies still focus on revealing the key issues and critical influencing factors of CCL (Arora et al., 2023; Kumar et al., 2023; Su et al., 2023); discussing how to achieve sustainable CCL using technologies such as the Internet of Things, digital twins, flexible sensing technologies, phase change materials, and robotic process automation (Huang et al., 2023; Lam & Tang, 2023; Li et al., 2024; Wu et al., 2023), and testing a multi-objective model of CCL’s sustainable performance through simulation methods (Abbas et al., 2023; Andoh & Yu, 2023; Habibur Rahman et al., 2023; Leng et al., 2024; Mirzaei et al., 2023; Navaei et al., 2023). The extant literature lacks empirical evidence on the relationship between sustainability innovation, supply chain resilience, and sustainability performance, especially in the Chinese CCL context. In addition, there is still controversy in existing literature regarding the relationship between the three dimensions of sustainability performance. Some argued that environmental and social performance contribute to economic performance (Chandra & Kumar, 2020; Mukherjee et al., 2022), while others deemed that economic performance underpins the achievement of environmental and social performance (Chowdhury & Hossan, 2014; Thong & Wong, 2018). Especially, Liao et al. (2023) found that China’s CCL industry pays more attention to economic sustainability, followed by environmental performance, and finally social performance, and recognised that further empirical research is needed. This provides an opportunity for our research.

At present, the consumption upgrade in China stimulates the rapid development of CCL (Su et al., 2023), but its overall level still lags behind that of developed countries (Peng et al., 2020). In order to reduce the disruption of CCL that still exists in the turbulent market environment (He & Liu, 2023), China’s CCLCs need to make efforts for sustainable development (General Office of China’s State Council, 2021). Therefore, it is significant to examine the impact of sustainability innovation and supply chain resilience on the sustainability performance of Chinese CCLCs. Based on the above motivation, this study aims to answer the following three research questions (RQs):

RQ1: Does sustainability innovation positively affect economic, environmental, and social performance in the context of CCL in China?

RQ2: Does supply chain resilience positively affect economic, environmental, and social performance in the context of CCL in China?

RQ3: What is the relationship between economic, environmental, and social performance in the context of CCL in China?

The originality of this study is to predict the impact of sustainability innovation and supply chain resilience collectively on sustainability performance and to reveal the internal links between economic, environmental, and social performance in the context of China’s CCL. This study can provide evidence for scholars to further expand the relationship between sustainability innovation, supply chain resilience, and sustainability performance and also offer important insights for managers of CCLCs to formulate sustainable development strategies. In the following sections, relevant hypotheses are developed in conjunction with the theoretical basis and tested using the Partial Least Squares Structural Equation Modelling (PLS-SEM) based on a survey of the Chinese CCLCs. Then, the empirical results, discussions, and implications are presented, followed by the conclusions, limitations, and future research directions.

2. Literature review and hypotheses development

2.1. Theoretical underpinning

The traditional resource-based view has been utilised to explain sustainability innovation and supply chain resilience (Ali et al., 2018; Tebaldi et al., 2018), but the static nature of resources in this theory struggles to explicate the dynamic evolving mechanisms of these strategies in volatile markets for business sustainability (Peng et al., 2020), which led to the application of Dynamic Capability Theory (DCT) in this study.

As the founder of DCT, Teece et al. (1997, p. 516) define dynamic capabilities as ‘the firm’s ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments’, which are usually formed by the three basic activities of sensing, seizing, and transforming (Teece, 2007). In a recent reflection, Teece, (2023) emphasise that ‘the dynamic capabilities framework can serve as a guide to empirical research’. This theory asserts that dynamic capabilities can effectively perceive changes in the internal and external environment and are the ability of managers to purposefully integrate and allocate internal and external resources, contributing to the long-term success of the organisation (Mishra et al., 2022; Wong et al., 2022). Hence, exposed to disruption risk (He & Liu, 2023; Le, 2023; Saari, 2023), CCLCs with dynamic capabilities are more likely to achieve business sustainability.

To achieve the Sustainable Development Goals (SDGs) that UN organisations are promoting, many academics have emphasised the significance of achieving sustainability performance in CCL (Abbas et al., 2023; Ali et al., 2023). As Turan and Ozturkoglu (2021) stated, the cold chain is a strategic tool to achieve social (ensuring human health by preventing and controlling the occurrence of hazardous food), economic (reducing costs), and ecological (reducing waste) sustainability. Hence, we consider the sustainability performance of CCL as the extent to which a CCLC can achieve objectives of economic, environmental, and social performance over time in a dynamic environment (Abeysekara et al., 2019; Le & Ikram, 2022). Su et al. (2023) claimed that innovation through investment in sustainable technologies, materials, processes, and business models would provide a continuous impetus for the sustainable transformation of CCLCs. According to DCT, sustainability innovation in this context can be seen as the ability of a CCLC to sense, identify, and seize innovation opportunities, adapt, and transform its resources (systems) in a risky business environment, which is a continuous, dynamic, and evolving process (Friedman & Ormiston, 2022). In addition, He and Liu (2023) looked into the dynamic process of supply chain resilience to protect the long-term growth of the agri-food CCL during times of resistance, adaptation, adjustment, and innovation due to unexpected public safety events. Khan and Ali (2023) agreed that supply chain resilience can safeguard the sustainable operation of the cold chain. From the perspective of DCT, supply chain resilience helps CCLCs perceive risks earlier, react quickly, and re-establish operations in a volatile and unpredictable environment, bringing them to a state of sustainability (Ali et al., 2022; Piprani et al., 2022; Sarkis, 2021; Umar & Wilson, 2024).

2.2. Sustainability innovation and sustainability performance

DCT asserts firms with dynamic capabilities are more likely to head towards sustainable business models (Bashir et al., 2022) and achieve operational excellence during periods of uncertainty (Mishra et al., 2022). Sustainability innovation, as a continuous dynamic capability, requires CCLCs to focus not only on economic benefits but also on reducing negative environmental and social impacts when investing in technology and innovating products (services), processes, facilities, and equipment (Su et al., 2023). In other words, firms with sustainability innovation would follow the Triple Bottom Line to promote dynamic innovation, which facilitates the fulfilment of stakeholders’ changing needs (Díaz-García et al., 2015). In this sense, the implementation of sustainability innovation in CCLCs contributes to improving sustainability performance.

Sustainability innovation is closely associated with the sustainable management of firms (Khan & Ponce, 2022). Jum’a et al. (2023) confirmed that supply chain innovation capability positively predicts sustainable supply chain performance. Product innovation, process innovation (Chandra & Kumar, 2020), technological innovation (Sundarakani & Onyia, 2021), market innovation (Chandra & Kumar, 2020), and business model innovation (Bashir et al., 2022; Zhao et al., 2022) have been revealed to be key factors in achieving organisational performance (Cuevas-Vargas et al., 2016). Besides, green innovation that lessens the detrimental environmental effects of economic activities can positively affect sustainability performance (Abubakar et al., 2022; Afum et al., 2023; Karman et al., 2024; Le & Ikram, 2022). By investigating 249 SMEs in Portugal and the UK, Almeida and Wasim (2023) revealed that eco-innovation in terms of products, processes, and green innovation are determinants of sustainability performance. On this basis, Afeltra et al. (2023) concluded that sustainability innovation is more capable of accomplishing sustainability performance in all three pillars compared to green innovation. Moreover, Elkhwesky et al. (2024) and Maletič et al. (2014) also confirmed that sustainability innovation positively affects economic, social, and environmental aspects of sustainability. An empirical investigation by Lee and Roh (2023) detected a positive correlation between open innovation and sustainability performance. Similarly, Farooq et al. (2024) affirmed that sustainability innovation contributes to environmental and financial sustainability. However, it has also been argued that sustainability innovation involves investment costs that may have a significant impact on economic performance and does not ensure success (Rodríguez-Espíndola et al., 2022; Souto, 2022; Wedari et al., 2023; Zhang et al., 2020). Furthermore, Dey et al. (2020) found that the link between sustainability innovation and environmental and social issues remains inconclusive.

In the CCL domain, Oriekhoe et al. (2024) verified in a comparative study of the United States and Africa that technological innovations can reform the food supply chain globally. Su et al. (2023) identified sustainable innovation as a key success factor for the food cold chain in China. For instance, Industry 4.0 (Kumar et al., 2024), IoT technology (Kumar et al., 2022b), and blockchain technology (Friedman & Ormiston, 2022) can facilitate information sharing, improve the visibility and traceability of the cold chain, and positively impact cold chain performance. That is to say, sustainable innovation solutions drive the sustainable development of CCL (Calati et al., 2023; Umate & Sawarkar, 2024). However, to the best of our knowledge, no existing studies tested the correlation between sustainability innovation and sustainability performance in the context of CCL in China. Considering that China is accelerating the modernization of CCL (Han et al., 2021), this study tends to argue that sustainability innovation can help Chinese CCLCs to gain more market share and also reduce environmental and social impacts. Therefore, we deduced that:

H1. Sustainability innovation positively affects economic performance in the Chinese cold chain logistics industry.

H2. Sustainability innovation positively affects environmental performance in the Chinese cold chain logistics industry.

H3. Sustainability innovation positively affects social performance in the Chinese cold chain logistics industry.

2.3. Supply chain resilience and sustainability performance

DCT views supply chain resilience as a dynamic capability that aids firms in successfully managing change in a dynamic manner to achieve sustained high performance (Teece et al., 1997; Yu et al., 2019). Supply chain resilience is often viewed as a ‘buffer’ or ‘lubricant’ for achieving sustainability performance and is designed to support firms in establishing sustainable development mechanisms in a dynamic environment. Without resilience, sustainability performance cannot be realised (Espiner et al., 2017). For example, firms adopting resilience initiatives such as redundant inventories and building trust and collaboration with partners tend to outperform those that do not, which also applies to the perishable product cold chain (Yavari & Zaker, 2020).

Many scholars have argued that supply chain resilience can improve firm performance or supply chain performance (Chowdhury & Quaddus, 2017; Mishra et al., 2022). Besides, Vargo and Seville (2010) found a measurable link between resilience and cash flow, return on investment, and profitability. Meanwhile, resilience strategies not only help to improve operational efficiency in the short term but also contribute to the sustainability of supply chains in the post-COVID-19 era (Sharma et al., 2022). Ruel and El Baz (2023) validated that supply chain resilience has a positive impact on financial performance during the COVID-19 outbreak. Xiao and Khan (2024) and Tian et al. (2024) testified that there is a positive correlation between supply chain resilience and supply chain performance in China’s healthcare industry and manufacturing industry, respectively. As for the relationship between supply chain resilience and sustainability performance, Chowdhury and Hossan (2014) verified that in the Bangladesh apparel industry, supply chain resilience positively influences environmental, economic, social, and operational sustainability. Rodríguez-González et al. (2023) conducted an empirical investigation of Mexican manufacturing industries and revealed that supply chain resilience significantly improves sustainability performance. Cui et al. (2023) observed that both active and passive supply chain resilience significantly and positively affect the three dimensions of sustainability performance in the Chinese manufacturing context. An empirical study of Vietnamese food SMEs by Le (2023) also indicated that supply chain resilience positively affects sustainable performance.

Furthermore, supply chain resilience can help CCL become more sustainable, which has been emphasised by Khan and Ali (2023), Taghikhah et al. (2023), Umar and Wilson (2024), and Gholami-Zanjani et al. (2021). However, in the post-COVID-19 era of the Chinese CCL industry, research on the relationship between supply chain resilience and sustainability performance is scarce. Based on the above discussion, the following hypotheses were proposed:

H4. Supply chain resilience positively affects economic performance in the Chinese cold chain logistics industry.

H5. Supply chain resilience positively affects environmental performance in theChinese cold chain logistics industry.

H6. Supply chain resilience positively affects social performance in the Chinesecold chain logistics industry.

2.4. Interrelationships between economic, environmental, and social performances

The three dimensions of sustainability performance, i.e. economic, environmental, and social performances, are believed to be interrelated. Chowdhury and Hossan (2014) verified a positive correlation between economic performance and social and environmental performance. Previous studies (Doane & MacGillivray, 2001; Thong & Wong, 2018) have reached similar conclusions. Generally, companies can only afford to take into account social issues if they are economically productive. For example, a company that is running at a loss may not pay or train employees in a timely manner, which can further cause instability or disruptions to the normal operation. Conversely, a profitable company would handle these issues easily. Similarly, if companies are economically unstable, they are unlikely to invest in environmental aspects, like the upgrading of energy-intensive equipment. Besides, improved environmental performance may have a positive impact on social performance (Henao & Sarache, 2022; Sajan et al., 2017). Environmental initiatives such as reducing energy consumption and waste emissions are conducive to improving the working environment and labour-management relations. On the contrary, ineffective environmental management may increase employees’ workload and affect their motivation.

However, Mukherjee et al. (2022) and Chandra and Kumar (2020) observed that environmental performance and social performance positively affect economic performance, whereas Gyamfi and Adebayo (2023) and Gupta and Racherla (2018) held that the relationship between environment and economic performance is bidirectional and positive. Differently, Wedari et al. (2023) asserted that firms with superior environmental performance may not be able to achieve the expected economic performance at the same time and contended that the impact of environmental performance on economic performance is delayed. Notably, Shakil et al. (2024) discovered that both environmental performance and social performance have a significant negative impact on financial performance through a survey of multinational transportation and logistics firms. Besides, Liang and Liu (2017) discovered that Chinese industrial enterprises currently pay little attention to environmental management, and concern for environmental performance does not significantly improve their economic performance. Also, Su et al. (2023) pointed out that the three dimensions of sustainability performance are still not balanced in the current Chinese CCL industry. In addition, Liao et al. (2023) revealed that the current Chinese CCLCs focus more on economic sustainability than environmental sustainability, and lastly, on social sustainability, which may be due to the high costs and low profit margins of the industry.

Besides, previous literature on the sustainability of CCL has mostly focused on discussing how to improve economic efficiency and reduce environmental impacts, with little attention to social performance (Liao et al., 2023), and there is also a lack of research on the relationship between the three dimensions. Although the future development of CCL must balance economic, environmental, and social aspects of sustainability (Mirzaei et al., 2023; Navaei et al., 2023; Ren et al., 2022), taking into account that in the scale expansion stage, managers of Chinese CCLCs may be more concerned with economic performance and cannot yet balance the relationship between the three dimensions (Liao et al., 2023; Su et al., 2023). Therefore, it is valuable to examine the relationship between economic, environmental, and social performance in the Chinese CCL context. Hence, the following hypotheses were developed:

H7. Economic performance has a positive impact on environmental performance in the Chinese cold chain logistics industry.

H8. Economic performance has a positive impact on social performance in the Chinese cold chain logistics industry.

H9. Environmental performance has a positive impact on social performance in the Chinese cold chain logistics industry.

Based on the discussion above, a research model was developed as shown in Figure 1, where sustainability innovation and supply chain resilience are jointly the antecedents of economic performance, environmental performance, and social performance; economic performance influences environmental performance and social performance, and environmental performance has a direct effect on social performance.

Figure 1. The conceptual framework.

3. Methodology

Following the positivist research paradigm, this study utilised a quantitative method with an explanatory and deductive design (Saunders et al., 2019). In addition, a questionnaire survey strategy was chosen to collect numeric data (Zikmund, 2003). This combined approach is often used to test the hypotheses between the research constructs to confirm the proposed theoretical framework (Rashid et al., 2024), which meets the research objectives of this study. The overall research process is shown in Figure 2.

Figure 2. Research process.

3.1. Measurement scales

When designing the research instrument of the study (see Figure 2), according to Gehlbach and Brinkworth (2011) and Boateng et al. (2018), it is reasonable to think of items created previously in the literature, which had been peer reviewed and validated, as inputs for new items or as ways to replicate them. Therefore, previously developed items were used for all of the constructs in this study. Of these, Sustainability Innovation (SI) borrowed 5 items from Souto (2022). Similarly, 6 items for Supply Chain Resilience (SCRE) were adopted from Yu et al. (2019). For the three dimensions of sustainability performance, 6 items for Environmental Performance (ENP) were borrowed from Souto (2022), 4 items for Economic Performance (ECP), and 5 items for Social Performance (SOP) were borrowed from Khan et al. (2021) (see Table 1). Furthermore, all measurements underwent slight revisions and modifications to align with the study’s context.

Table 1. Measurement items.

Construct	Revised items	Sources
Sustainability Innovation (SI)	1. Our company focuses on significant improvements in products (services), processes, and/or organisational methods oriented towards increasing sustainability.	Souto (2022)
	2. Our company encourages radical innovation (breakthrough) oriented towards increasing sustainability.
	3. Our company promotes business model innovation and/or business concept innovation with an important contribution towards increasing sustainability.
	4. Our company endeavours to improve sustainable innovativeness and sustainable innovation capacity.
	5. Our company supports open sustainability innovation.
Supply Chain Resilience (SCRE)	1. Our company with its supply chain partners are well prepared to deal with the financial outcomes of supply chain risk events.	Yu et al., (2019)
	2. Our company with its supply chain partners are able to adequately respond to unexpected risks by quickly restoring its product flow.
	3. Our company with its supply chain partners can quickly return to its original state after experiencing risky events.
	4. Our company with its supply chain partners can move to a new, more desirable state after experiencing risky events.
	5. Our company with its supply chain partners have the ability to maintain a desired level of control over the structure and function of the supply chain during the time of risky events.
	6. Our company with its supply chain partners have the ability to extract meaningful and useful knowledge from risky events.
Economic Performance (ECP)	1. Our sales have increased over the past three years.	Khan et al. (2021)
	2. Our profit has grown over the past three years.
	3. Our return on investment has increased over the past three years.
	4. Our market share has increased over the past three years.
Environmental Performance (ENP)	1. We have improved the company’s environmental situation over the past three years.	Souto (2022)
	2. We have reduced air emissions across our processes over the past three years.
	3. We have reduced hazardous/harmful/toxic material consumption over the past three years.
	4. We have increased the use of renewable energy and recycled materials over the past three years.
	5. We have reduced waste across our processes over the past three years.
	6. We have reduced resource consumption (energy, raw materials, etc.) over the past three years.
Social Performance (SOP)	1. Employees’ occupational health and safety have improved over the past three years.	Khan et al., (2021)
	2. Employees’ training and education have increased over the past three years.
	3. Our employee motivation has increased over the past three years.
	4. Our employee job satisfaction has improved over the past three years.
	5. Our employee turnover ratio has decreased over the past three years.

As for scale points, Stem and Noazin (1985) investigated that 5- and 7-point Likert scales are the most reliable and that reliability decreases significantly when more than 7 points are used. Generally speaking, the 7-point Likert scale provides more options than the 5-point Likert scale and can be used to gather more detailed insights from the respondents on the items. Kusmaryono et al. (2022) and Cox (1980) highlighted that the 7-point Likert scale is the most effective in terms of reliability and validity. Hence, a 7-point Likert scale from strongly disagree (1) to strongly agree (7) was employed in the study in order to obtain more detailed feedback from the respondents on the constructs, especially economic, environmental, and social performance.

To validate the questionnaire further, face validity and content validity should be checked (Cavana et al., 2001; Mustapha et al., 2023). By distributing the draft questionnaire to three university professors and two senior managers in the Chinese cold chain logistics sector, the content validity of the scales was assessed to ensure their adequacy and representativeness (Cavana et al., 2001). Subsequently, 35 Chinese CCLC managers were chosen to evaluate face validity. Their suggestions were incorporated to make sure the questionnaire was clear, readable, and understandable.

3.2. Population, sampling design and sample size

This study focused on managers of Chinese CCLCs with 3 years of industry experience. The cold chain logistics industry was selected because of the rapid increase in Chinese market demand for cold chain products like meat, fruits, vegetables, fish, vaccines, and pharmaceuticals. As China becomes more concerned about food safety and low-carbon development, CCLCs are accelerating innovation and building resilience to offer sustainable high-quality services (General Office of China’s State Council, 2021). The choice of managers as respondents was based on their expertise in operations management and their broader understanding of economic, environmental, and social concerns (Le & Ikram, 2022). Furthermore, to guarantee that the respondents were familiar with the business operation and their company’s performance, 3 years of industry experience was required, which was met by the screening questions on the questionnaire’s cover page.

The study utilised a non-probability judgmental sampling technique due to the unavailability of a comprehensive database of Chinese CCLCs and the difficulty in obtaining a list of their managers (Mohammad & Quoquab, 2024; Perla & Provost, 2012). As Memon et al. (2020) highlighted that organisational-level research may have a smaller sample size than individual-level research, two approaches were adopted for the minimum sample size. First, following Boomsma’s (1982), we obtained the required sample size n using the formula (n ≥ 50r² – 450r + 1100), where r is the ratio of indicators to latent variables. As r = 5.2 in our study, the required minimum sample size is 112. Second, G*Power software (version 3.1.9.7) was used to calculate the minimum sample size (Hair et al., 2022). By setting the effect size (f²) of 0.15, the error probability (α) of 0.05, the power level (1-β) of 0.8, and the predictor number of 4, the recommended sample size is 85. Taken together, a sample size of 112 can yield reliable conclusions.

3.3. Data collection and profile of the respondents

An online questionnaire designed on the Wenjuanxing platform (www.wjx.cn) was utilized to collect data in this study. The questionnaire began with a cover page describing the survey’s purpose and target population, emphasising the voluntary and anonymous nature of participation, and ensuring that the data would be kept confidential and used only for academic purposes. The remaining two sections collected the respondents’ opinions on all constructs as well as demographic characteristics. Through WeChat and e-mail, the questionnaire was distributed to 300 managers from different Chinese CCLCs between August and September 2023. Finally, 204 valid responses were obtained, yielding an effective response rate of 68%.

Table 2 shows the demographic characteristics of the participants. Among them, by gender, 63.7% of the respondents were male. The 36–45 age group has the highest proportion of 43.6%. 61.2% of them have a bachelor’s or master’s degree and above. Most respondents are business operation managers, and transportation and warehousing are the main businesses. The employee numbers indicate that most companies surveyed have less than 200 employees, and industry subsectors are mainly in ‘fruits and vegetables’ and ‘meat, aquatic, and dairy products’.

Table 2. Demographic profile.

Title	Description	Frequency	Percentage (%)
Gender	Female	74	36.3
	Male	130	63.7
Age	18–25	1	0.5
	26–35	61	29.9
	36–45	89	43.6
	46–55	34	16.7
	>55	19	9.3
Education	High school or below	25	12.3
	College degree	54	26.5
	Bachelor	89	43.6
	Master or above	36	17.6
Position	Executive	49	24.0
	Business operation managers	87	42.6
	Supply chain manager	39	19.1
	Risk managers	18	8.8
	Other managers	11	5.4
Main business fields	Warehousing	53	26.0
	Transportation	65	31.9
	Distribution	32	15.7
	Information transaction	21	10.3
	Integrated business	33	16.2
Number of employees	<50	24	11.8
	51–100	60	29.4
	101–200	85	41.7
	>200	35	17.2
Type of sub-industry	Fruits and vegetables	61	29.9
	Meat, aquatic and dairy products	96	47.1
	Vaccines, drugs and special products	17	8.3
	Any two or more of the above types involved	30	14.7

3.4. Data analysis approach

This study used the PLS-SEM with SmartPLS 4 software (version 4.0.9.5) (Ringle et al., 2024) to test the theoretical model. The PLS-SEM approach can predict the endogenous construct, maximise its variance, and offer simultaneous explanation and prediction assessment of the model (Hair et al., 2019), fulfilling the study’s aim. The two-stage approach of PLS-SEM entails the estimation of both the measurement model and the structural model (Hair et al., 2022). In particular, after assessing the measurement model’s validity and reliability, the structural model would examine the interconnections between the research constructs and assess the model’s explanatory and predictive power.

4. Results

4.1. Common method bias

The Common Method Bias (CMB) from self-reporting may have a substantial impact on the relationship between different constructs due to that the same respondents offered measures of exogenous and endogenous variables (Podsakoff et al., 2003). To ensure that CMB does not affect this study, procedural and statistical methods were applied. In the procedural approach, we emphasised the anonymous and voluntary nature of the survey and the clarity of constructs and items. As for the statistical method, Harman’s single-factor test (Podsakoff et al., 2003) and the VIF threshold (Kock, 2015) were recommended to diagnose CMB. Using SPSS 27.0 software to perform Harman’s single-factor test, as shown in Table 3, the results demonstrated that the first factor’s explanation of the total variance is only 38.278%, which is less than the 50% threshold that is typically linked to method bias. Then, all VIFs in the outer model of the present study are less than 3.3. In combination, CMB is not a significant issue for this study (Harman, 1976; Kock, 2015).

Table 3. Total variance explained.

	Initial Eigenvalues			Extraction sums of squared loadings
Component	Total	% of Variance	Cumulative %	Total	% of Variance	Cumulative %
1	9.952	38.278	38.278	9.952	38.278	38.278
2	2.687	10.333	48.612	2.687	10.333	48.612
3	2.217	8.526	57.138	2.217	8.526	57.138
4	2.049	7.879	65.017	2.049	7.879	65.017
5	1.637	6.296	71.314	1.637	6.296	71.314
6	0.642	2.467	73.781
7	0.567	2.179	75.96
8	0.548	2.106	78.067
9	0.485	1.866	79.933
10	0.471	1.81	81.743
11	0.42	1.615	83.358
12	0.412	1.586	84.944
13	0.383	1.474	86.418
14	0.377	1.449	87.868
15	0.337	1.295	89.162
16	0.33	1.27	90.433
17	0.321	1.236	91.669
18	0.309	1.189	92.858
19	0.295	1.134	93.991
20	0.278	1.071	95.062
21	0.257	0.989	96.051
22	0.239	0.919	96.97
23	0.233	0.897	97.867
24	0.206	0.791	98.658
25	0.203	0.782	99.441
26	0.145	0.559	100

Extraction Method: Principal Component Analysis.

4.2. Measurement model assessment

The reliability and validity of the outer model should be confirmed before assessing the structural model (Hair et al., 2022). In Table 4, the values of out loadings, Cronbach’s α (CA), composite reliability (rho_a), and composite reliability (rho_c) are all larger than 0.70, ensuring reliability at the item and construct levels (Hair et al., 2019). Also, Table 4 demonstrates that convergent validity was ensured based on the values of AVE that exceeded 0.5 for all constructs (Fornell & Larcker, 1981).

Table 4. Assessment of measurement model.

Constructs	Items	Loading	CA	rho_a	rho_c	AVE
SI	SI1	0.847	0.901	0.906	0.926	0.716
	SI2	0.827
	SI3	0.824
	SI4	0.868
	SI5	0.863
SCRE	SCRE1	0.846	0.902	0.905	0.924	0.67
	SCRE2	0.814
	SCRE3	0.819
	SCRE4	0.826
	SCRE5	0.829
	SCRE6	0.776
ECP	ECP1	0.869	0.891	0.891	0.924	0.754
	ECP2	0.885
	ECP3	0.875
	ECP4	0.843
ENP	ENP1	0.825	0.909	0.910	0.930	0.688
	ENP2	0.811
	ENP3	0.831
	ENP4	0.852
	ENP5	0.846
	ENP6	0.809
SOP	SOP1	0.845	0.904	0.904	0.928	0.722
	SOP2	0.848
	SOP3	0.867
	SOP4	0.834
	SOP5	0.855

Next, discriminate validity was tested based on the Fornell-Larcker criterion and the heterotrait-monotrait (HTMT) methods. From Table 5, the square root of the AVE for all constructs on the diagonal is the highest in raw and column, compared to the correlation coefficients with the other constructs. In Table 6, the HTMT values for all constructs are all less than the threshold of 0.85. Based on these results, discriminant validity was established.

Table 5. Fornell-Larcker test.

	SI	SCRE	ECP	ENP	SOP
SI	0.846
SCRE	0.403	0.819
ECP	0.386	0.402	0.868
ENP	0.442	0.499	0.422	0.829
SOP	0.468	0.44	0.454	0.342	0.85

Table 6. HTMT method.

	ECP	ENP	SCRE	SI	SOP
ECP
ENP	0.467
SCRE	0.442	0.55
SI	0.426	0.485	0.447
SOP	0.506	0.376	0.483	0.514

4.3. Structural model evaluation

The structural model assessment involved collinearity issues, path coefficients, the effect sizes (f²), coefficients of determination (R²), and predictive relevance (Q²) (Hair et al., 2022). Hypothesis testing was performed with the aid of bootstrapping method (5000 subsamples, two-tailed, significance level = 0.05) (Henseler et al., 2015; Zhang et al., 2024).

Table 7 shows that all of the variance inflation factors (VIF) for the inner model are less than 3, confirming the absence of collinearity (Hair et al., 2019). As shown in Table 7 and Figure 3, the estimated coefficient of β = 0.267 (t = 3.667, p < 0.001) for the relationship between SI and ECP is positive, indicating strong support for H1. Also, the estimated coefficient of β = 0.233 (t = 4.334, p < 0.001) for the relationship between SI and ENP significantly supports H2, and the estimated coefficient of β = 0.279 (t = 4.093, p < 0.001) for the relationship between SI and SOP significantly supports H3. Similarly, the proposed connections between SCRE and ECP, ENP, and SOP are backed up by the estimated coefficients of β = 0.295 (t = 4.600, p < 0.001), β = 0.323 (t = 4.813, p < 0.001), and β = 0.225 (t = 3.236, p < 0.01) (see Table 7 and Figure 3). The estimated coefficient of β = 0.202 (t = 3.136, p < 0.01) for the relationship between ECP and ENP is significant, representing strong support for H7. Likewise, as shown in Table 7 and Figure 3, H8, ECP enhances SOP, which is supported by an estimated coefficient of β = 0.257 (t = 3.920, p < 0.001). However, surprisingly, the result shows the impact of ENP on SOP is insignificant with an estimate coefficient β = −0.002 (t = 0.021, p > 0.05). In general, as presented in Figure 3, all hypotheses are confirmed except for H9 (ENP->SOP).

Figure 3. Research framework with results.

Notes: SI, sustainability innovation; SCRE, supply chain resilience; ECP, economic performance; ENP, environmental performance; SOP, social performance.

Table 7. Hypotheses testing.

Hypotheses	Relationships	β	SE	VIF	R^2	f^2	Q^2	t-values	Decision
H1	SI -> ECP	0.267***	0.062	1.194	0.222	0.077	0.160	4.334	Supported
H2	SI -> ENP	0.233***	0.064	1.286	0.350	0.065	0.233	3.667	Supported
H3	SI -> SOP	0.279***	0.068	1.370	0.345	0.087	0.243	4.093	Supported
H4	SCRE -> ECP	0.295***	0.064	1.194		0.093		4.600	Supported
H5	SCRE -> ENP	0.323***	0.067	1.306		0.123		4.813	Supported
H6	SCRE -> SOP	0.225**	0.069	1.467		0.053		3.236	Supported
H7	ECP -> ENP	0.202**	0.064	1.285		0.049		3.136	Supported
H8	ECP -> SOP	0.257***	0.066	1.348		0.075		3.920	Supported
H9	ENP -> SOP	−0.002	0.075	1.538		0.000		0.021	Not supported

Notes: SI, sustainability innovation; SCRE, supply chain resilience; ECP, economic performance; ENP, environmental performance; SOP, social performance; *p-value < 0.05, **p-value < 0.01, ***p-value < 0.001.

Next, f², R², and Q² were estimated, with the results presented in Table 7. According to Cohen (1988), SI (f² = 0.087) has a stronger influence on SOP than SCRE (f² = 0.053), while SCRE has a greater impact on ECP (f² = 0.093) and ENP (f² = 0.123) than SI (f² = 0.077 and 0.065, respectively). In explaining the variance in SOP, there is a weak effect of ECP (f² = 0.075) and no effect of ENP (f² = 0.000). Compared to the recommendations of R² suggested by Cohen (1988), the model explains the variance of 22.2% for ECP, 35% for ENP, and 34.5% for SOP, showing good explanatory power. Besides, all Q² values of endogenous constructs are greater than 0, supporting the model’s predictive relevance (Hair et al., 2019).

To draw conclusions that are meaningful to management practices, out-of-sample predictive power should be tested through the PLSpredict algorithm (Hair et al., 2019). The PLS-SEM analysis indicates that only two indicators make larger prediction errors than the naive LM benchmark, indicating a medium predictive power (Table 8).

Table 8. PLSpredict.

	Q²predict	PLS-SEM_RMSE	LM_RMSE
ECP1	0.160	1.551	1.583
ECP2	0.163	1.583	1.572
ECP3	0.127	1.635	1.615
ECP4	0.169	1.551	1.526
ENP1	0.215	1.491	1.542
ENP2	0.225	1.441	1.493
ENP3	0.221	1.493	1.538
ENP4	0.227	1.434	1.505
ENP5	0.214	1.487	1.557
ENP6	0.155	1.541	1.599
SOP1	0.193	1.506	1.538
SOP2	0.200	1.451	1.472
SOP3	0.211	1.430	1.474
SOP4	0.208	1.443	1.480
SOP5	0.189	1.551	1.600

5. Discussion and implication

The analysed results above provide strong support for the discussion of research constructs’ relationships. SI positively affects ECP, ENP, and SOP (H1, H2, and H3). The findings are in line with prior studies (Abubakar et al., 2022; Elkhwesky et al., 2024; Farooq et al., 2024; Lee & Roh, 2023; Le & Ikram, 2022), which suggested that enhancing the sustainability innovation capability of a CCLC contributes to sustainability performance. In addition, the findings are in line with the DCT, which contends that through sensing the environment, capturing opportunities, or re-configuring organisational resources and capabilities (Bashir et al., 2022; Teece, 2007), sustainability innovation in different aspects of CCL services, processes, and business models can improve profitability, reduce environmental impacts, enhance the well-being of employees, and ultimately achieve sustainability performance.

SCRE positively influences ECP, ENP, and SOP (H4, H5, and H6). The findings are consistent with the previous literature (Chowdhury & Hossan, 2014; Cui et al., 2023; Le, 2023; Mishra et al., 2022; Rodríguez-González et al., 2023; Sharma et al., 2022). The results reveal that by developing supply chain resilience, CCLCs in a dynamic environment can strengthen cooperation with supply chain partners and enhance their awareness of risk prediction (preparedness). When they encounter a risk event, they have the capability to deploy the resources of the supply chain to respond in a timely manner and maintain the operation of their cold chain business, while the business will be quickly restored to a normal or even better state. Thus, this is also in line with the DCT (Piprani et al., 2022; Teece, 2007). In other words, CCLCs with supply chain resilience can retain stable economic returns, are more likely to realise reductions in resource consumption and waste emissions, and increase the confidence of their employees, thereby promoting their sustainable development.

This study examined the interrelationships among the three dimensions of sustainability performance (H7, H8, and H9). The results display that ECP has a positive effect on ENP and SOP, which is consistent with the findings of Chowdhury and Hossan (2014), Thong and Wong (2018), and Doane and MacGillivray (2001). For current CCLCs, economic sustainability is the basis for realising social and environmental sustainability. Besides, it shows an insignificant relationship between ENP and SOP, which suggests that the realisation of ENP does not guarantee SOP, indicating that economic sustainability may be a better guarantee of social sustainability than environmental sustainability.

5.1. Theoretical implication

This study contributes positively to the theoretical literature on dynamic capability strategies and sustainability in different ways.

First, with the difference from Eikelenboom and de Jong (2019), who investigated the impact of integrated dynamic capabilities on sustainability performance but lacked theoretical underpinnings, this study explained the mechanism by which SI and SCRE were transformed into sustainability performance under the dynamic capabilities framework, extending the application of DCT.

Second, this study demonstrated the significance of integrating SI and SCRE strategies in a volatile environment, bridging a gap in the literature. In particular, SI emphasises open and comprehensive sustainability-oriented innovation within the firm, while SCRE focuses on collaboration with supply chain partners externally in order to consolidate resources and enhance resistance to risk.

Third, although previous studies have provided different results that the three pillars of sustainability performance are related to each other, the present study emphasises the significant positive impact of economic performance on the other two pillars, which reinforces one of the research streams that economic performance is a prerequisite for improving environmental and social performance (Doane & MacGillivray, 2001; Thong & Wong, 2018). The insignificant effect of environmental performance on social performance suggests that at this stage improving the economic performance is a critical path to solving social issues such as employee satisfaction and motivation, which is in line with Liao et al. (2023).

Last but not least, the research context of this study is unique. As the CCL does not allow breakage in the chain, has special requirements for supply chain collaboration, and relies on technology, the results are potentially valuable.

5.2. Practical implication

This study provides important practical implications for policymakers and practitioners.

For practitioners, first, the insight that developing multiple dynamic capabilities contributes to sustainability suggests that they should avoid using a single strategy and deploy both sustainability innovation and supply chain resilience as part of their strategic plan. Also, in today’s increasingly competitive world, managers must understand the highly dynamic and complex nature of sustainability innovation and supply chain resilience, continuously assess the environment, seize opportunities, enhance the integration and utilisation of internal and external resources, and build dynamic development mechanisms. Second, since improvements in economic aspects may support enhancements in environmental and social aspects by reducing energy consumption, increasing efficiency, and providing social value, practitioners should prioritise securing economic sustainability at this stage.

For policymakers, first, attention should be paid to CCLCs that are actively implementing innovation and resilience, as they are likely to be more concerned with sustainability, which would reduce the environmental and social burden of the government to some extent. Second, policymakers should launch regulatory policies in a timely manner, clarify industry standards, and guide CCLCs to speed up the renewal of old facilities and equipment and the application of energy-saving and environmentally friendly technologies. Third, considering the substantial initial investment in cold chain infrastructure (Han et al., 2021; Kumar et al., 2024), policymakers should establish subsidies to relieve CCLCs’ financial burdens. Furthermore, policymakers may also provide support to CCLCs in terms of personnel training and industrial cooperation, thereby promoting the holistic development of this industry.

In addition, the pathway towards sustainable CCL identified in this study can provide lessons for other developing countries, which is essential for reducing global food losses and energy consumption as well as addressing social issues such as global hunger and health.

6. Conclusions, limitations, and future research directions

In this paper, we hypothesise sustainability innovation and supply chain resilience as important strategic tools to achieve sustainability in the Chinese CCL context. The empirical results reveal the positive effects of sustainability innovation and supply chain resilience on sustainability performance, as well as the positive impacts of economic performance on environmental and social performances, and emphasise that environmental performance has a limited impact on social performance. Theoretically, this study explains the relationship between sustainability innovation as well as supply chain resilience and sustainability performance under the dynamic capabilities framework. It combines these two dynamic capabilities into a single model to predict how they relate to sustainability performance and confirms that improving economic performance is needed in the Chinese CCL industry in order to improve environmental and social performance. In addition, it provides significant implications for practitioners and policymakers. For example, CCLCs can foster sustainability-oriented innovations and enhance collaboration with supply chain partners to develop continuous innovative and resilience capacity and focus on achieving economic benefits, which are not only essential for their survival but also the basis for further environmental and social performances. Policymakers should introduce regulatory and subsidy policies to encourage practitioners to develop sustainability-oriented innovations and supply chain resilience.

It is undeniable that this study still has limitations that should be further addressed. First, this cross-sectional study was limited in identifying influential relationships between constructs in a dynamic environment, so performing a longitudinal study could be considered in the future. Second, data were collected from Chinese CCLCs, and future research could be transferred to other countries or industries. Third, in the research model, considering supply chain resilience (Abeysekara et al., 2019) and sustainability performance (Khan et al., 2021) as hierarchical constructs could be an interesting research direction.

Author contributions

Baohua Zhang: Writing – original draft, Conceptualization and design, Methodology, Data analysis, Writing – review & editing. Jihad Mohammad: Supervision, Methodology, Data analysis, Writing – review & editing. Both Baohua Zhang and Jihad Mohammad contributed to the final version of the manuscript and agreed to be accountable for all aspects of the work.

Acknowledgements

This research has no conflicts of interest to declare. Any errors are our own.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Additional information

Funding

This research is supported by the Henan Soft Science Research Program Project in 2023 under Grant No. 232400411078, and the Henan Soft Science Research Program Project in 2024 under Grant No. 242400411184.

Notes on contributors

Baohua Zhang

Baohua Zhang (Mr) is a PhD student in the Business and Management Program, UCSI Graduate Business School, UCSI University, Kuala Lumpur, Malaysia. He has received his Master degree in Logistics Engineering and Management from Dalian Maritime University in 2010, and is also the head of modern logistics management major at Henan Vocational College of Water Conservancy and Environment. His current research focus is sustainable cold chain logistics.

Jihad Mohammad

Jihad Mohammad (Dr) is an Assistant Professor at UCSI Graduate Business School, UCSI University, Kuala Lumpur, Malaysia. He has received his DBA degree from Universiti Kebangsaan Malaysia. He has published several articles in peer-reviewed international journals. He has versatile career exposure. He has conducted several workshops for postgraduate students regarding research methodology and Structural Equation Modelling. His area of research interest includes organizational citizenship behavior, psychological ownership, psychological capital, leadership, innovation, and Islamic work ethics.