Do Supply Chain Management, ESG Sustainability Practices, and ICT Have an Impact on Environmental Sustainability?

Abstract

Can supply chain strategies, ESG practices, and digital innovations be the game-changers the planet needs for a sustainable future? Motivated by this question, this study investigates the drivers of CO₂ emissions, focusing on supply chain management (GSC), ESG sustainability practices, and Information and Communication Technology (ICT) in China from 2002Q4 to 2024Q4. Utilizing a series of wavelet tools—including wavelet coherence (WTC), partial wavelet coherence (PWC), and multiple wavelet coherence (MWC)—the study uncovers associations across time and frequency domains. To the best of the authors’ knowledge, this is the first study to examine these dynamics within the Chinese context using advanced wavelet techniques. The WTC results reveal that GSC, ICT, and patents are positively associated with CO₂ emissions, particularly during 2008–2016 and 2018–2024, while ESG practices reduced emissions before 2016 but became positively linked to emissions afterward. MWC and PWC analyses confirm that these drivers influence CO₂ within 1–4-year bands, while wavelet Granger causality tests indicate weak short-term but strong medium- to long-term causal relationships among ESG, GSC, PAT, ICT, and CO₂ emissions. Based on these results, policy recommendations are formulated.

1. Introduction

China remains the world’s largest CO₂ emitter, producing approximately 11.3 Gt of CO₂ in 2021, nearly a third of the global total, with about 10.5 Gt from fossil fuel combustion [1]. However, recent progress is encouraging: clean power expansion drove a 1.6% year-on-year drop in emissions (Q1 2025) and a 1% decline over the past 12 months [1]. Although the United States remains the largest emitter on a per capita basis, China overtook the U.S. in total greenhouse gas emissions around 2006, marking a significant shift in the global emissions landscape. Coal’s dominance—responsible for 79% of energy-sector CO₂ in 2022—has been challenged by a surge in renewables, with wind, solar, hydro, and nuclear now accounting for 44% of electricity supply as of May 2024. Between 2000 and 2020, CO₂ emissions grew by ~5.4% annually compared to GDP growth of ~9.5%, showing a gradual improvement in carbon intensity. China’s dual “30–60” commitment—peaking emissions by 2030 and reaching carbon neutrality by 2060—frames its policy horizon [1]. Under its “1 + N” climate policy, the country aims to boost non-fossil energy to 25% by 2030 and expand total renewable capacity to ~1200 GW by 2030. As of 2024, it surpassed this target early with over 1400 GW of wind and solar capacity. Policy tools include a national carbon trading scheme, Five-Year Plan targets for energy and carbon intensity reductions, and massive investments in clean energy tech—making China the world’s top investor in renewables, pouring USD 546 billion in 2022. Still, challenges remain: coal-fired generation persists, with approvals totaling 94.5 GW in 2024, prompting calls to accelerate the energy transition [2].

ESG sustainability is a pivotal driver in advancing China’s low-carbon transition by promoting environmental accountability, fostering cleaner production, and guiding capital toward sustainable investments, thereby exerting a direct downward pressure on CO₂ emissions [3]. By embedding environmental considerations within corporate strategies and financial decisions, ESG frameworks incentivize firms to adopt energy-efficient technologies, reduce pollution, and transparently report emissions, aligning market behaviors with China’s carbon neutrality goals [4]. Empirical evidence demonstrates that firms with strong ESG practices exhibit significantly lower CO₂ emissions, as their commitment to environmental stewardship translates into operational changes that curb fossil fuel dependence [5,6]. Moreover, ESG factors influence capital allocation, with investors increasingly directing funds toward firms with superior ESG ratings, thereby creating a market-based mechanism that pressures high-emission industries to decarbonize to maintain competitiveness [7]. In China, initiatives such as mandatory environmental information disclosure and green credit policies reinforce ESG’s role as a complementary tool to regulatory frameworks, ensuring that environmental sustainability commitments are integrated into the economic fabric. However, the effectiveness of ESG in driving down CO₂ emissions is contingent upon robust governance and enforcement, highlighting the need for continuous policy refinement to ensure ESG practices are not only adopted in form but are impactful in reducing emissions and supporting sustainable economic growth [8,9].

Global supply chain management (GSC) plays a critical role in determining CO₂ emissions by shaping production patterns, transportation intensity, and resource allocation across interconnected economies, thereby acting as both a driver and potential mitigator of environmental impacts [10]. Efficient GSC practices can significantly reduce CO₂ emissions by optimizing logistics, minimizing transportation distances, reducing inventory-related energy consumption, and integrating cleaner technologies across supplier networks [11]. Conversely, fragmented and complex supply chains increase emissions through energy-intensive manufacturing processes, excessive transportation, and reliance on carbon-intensive logistics, particularly in economies heavily integrated into global trade like China. Empirical evidence indicates a strong positive relationship between GSC expansion and CO₂ emissions, as globalized production networks often shift high-emission manufacturing to developing economies, contributing to rising carbon footprints [12]. However, GSC also presents opportunities for emission reductions through the adoption of green supply chain initiatives, including low-carbon logistics, supplier decarbonization incentives, and circular economy practices, aligning supply chain management with global climate targets [13,14]. In the context of China, which is central to global supply chains, optimizing GSC practices is essential for reducing CO₂ emissions while maintaining economic competitiveness, emphasizing the need for coordinated policy and private sector strategies that integrate environmental sustainability into supply chain decisions.

Information and Communication Technology (ICT) serves as a significant determinant of CO₂ emissions, exerting both direct and indirect influences on environmental sustainability depending on the scale, efficiency, and application of digital technologies [15,16]. On one hand, the expansion of ICT infrastructure—such as data centers, network systems, and electronic device production—contributes directly to energy consumption and associated CO₂ emissions, particularly in economies with carbon-intensive energy structures like China [17,18]. On the other hand, ICT plays a pivotal role in reducing CO₂ emissions indirectly by enabling energy-efficient practices, optimizing industrial processes, facilitating digital monitoring of emissions, and supporting the transition to smart grids and intelligent transportation systems [19]. Empirical evidence supports this dual relationship, showing that while ICT expansion initially increases emissions, the deployment of advanced ICT solutions contributes to decoupling economic growth from carbon emissions in the long term [20]. In China, the integration of ICT into manufacturing and energy sectors has facilitated emission monitoring and process optimization, contributing to the country’s broader carbon neutrality targets, even as the rapid digital transformation has led to increased energy demand in the ICT sector itself [21]. This complex relationship underscores the critical need for policies that enhance the role of ICT as an enabler of decarbonization while managing its direct carbon footprint, positioning ICT as both a challenge and a solution in China’s low-carbon development strategy.

Building on the above discussion, this study investigates the drivers of CO₂ emissions by addressing the following research questions:

❖

Does global supply chain management impact CO₂ emissions in China?

❖

Does Information and Communication Technology influence China’s CO₂ emissions?

❖

Does ESG sustainability impact China’s CO₂ emissions?

❖

Does patent innovations impact China’s CO₂ emissions?

Contribution of Study

China, as a global economic powerhouse, stands at the forefront of the climate challenge. With CO₂ emissions threatening climate goals and SDG progress, China’s commitment to carbon neutrality by 2060 demands transformative strategies. This study dives into how ESG sustainability practices, ICT development, GSC management, and patent innovations (PAT) shape CO₂ emissions within China’s rapidly evolving landscape. Unpacking the roles of these critical drivers is key to steering China toward a low-carbon future while sustaining economic momentum. By exploring these interconnected factors, this research offers fresh, actionable insights for crafting effective policies to address the dual challenge of emissions reduction and sustainable development in China.

From a theoretical lens, this study injects new energy into the sustainability discourse by bridging gaps in the literature on China’s climate efforts. Unlike fragmented analyses, it brings ESG, ICT, GSC, and PAT under one framework, revealing how these elements interact to influence emissions outcomes. This integrated perspective underscores the need to align ESG frameworks with digital transitions, green supply chain management, and patent-driven clean innovations to advance China’s carbon neutrality agenda. By addressing these factors collectively, this study offers a roadmap for transforming policy guidelines from isolated actions into comprehensive, adaptive frameworks essential for China’s SDG journey.

On the practical front, this research uncovers clear, targeted policy pathways for reducing CO₂ emissions without compromising growth. It calls for ramping up ESG compliance within supply chains, leveraging ICT to drive energy efficiency, and incentivizing clean technology patents to catalyze industrial decarbonization. Strengthening the synergy between ESG practices and ICT-enabled green supply chains, while accelerating green innovation through patent systems, is crucial for propelling China toward its climate goals. These actionable insights are vital for advancing SDGs 7, 9, 12, and 13.

Methodologically, this study sets itself apart by employing advanced wavelet-based tools that capture the pulse of dynamic relationships across time and frequency. By applying wavelet coherence, partial wavelet coherence, and multiple wavelet coherence, the study reveals how ESG, ICT, GSC, and PAT interact with CO₂ emissions over short-, medium-, and long-term horizons. This high-resolution lens enables a granular understanding of when and how these factors impact emissions, offering policymakers and practitioners robust, time-sensitive evidence to design adaptive, integrated climate strategies. Through this approach, the study not only enriches the empirical literature but also equips China with sharper tools to navigate its path toward a greener, more resilient future.

The next sections are as follows: Section 2 discusses the theoretical framework and literature review; Section 3 presents the data and methods; Section 4 displays the findings; and Section 5 provides the conclusion and policy implications.

2. Theoretical Framework and Literature Review

2.1. Theoretical Framework

Environmental, social, and governance (ESG) practices are increasingly recognized as crucial mechanisms for enhancing firms’ environmental performance. By promoting cleaner production processes, efficient resource utilization, and strict compliance with environmental regulations, ESG initiatives can play a pivotal role in curbing CO₂ emissions [4]. Simultaneously, the development of Information and Communication Technology (ICT) facilitates digitalization through automated systems, real-time monitoring, and data-driven energy management, all of which contribute to operational efficiency and carbon footprint reduction across various sectors [3,17,18]. However, this positive contribution is complicated by the potential rebound effect—where energy savings from ICT adoption are offset by increased energy demand due to higher usage and production [22,23]. Thus, ESG and ICT function as key channels through which firms adapt their production structures, energy consumption patterns, and reporting transparency, ultimately influencing their trajectory toward environmental sustainability.

In parallel, global supply chain (GSC) management exerts a significant influence on environmental outcomes, particularly CO₂ emissions, through its integral role in logistics, transportation, and cross-border trade. These activities are often energy-intensive and contribute substantially to global emissions. However, when GSC practices integrate ESG principles—such as route optimization, low-carbon transportation, and sustainable procurement—they can meaningfully reduce emissions and promote greener supply chains [24,25]. Furthermore, patent innovations (PAT) serve as proxies for technological advancement and reflect a firm’s capacity to improve production efficiency and reduce ecological harm. While clean technology patents are generally associated with emissions mitigation, the environmental effects of other forms of innovation depend on the industry context and scale of implementation [26,27].

2.2. Literature Review

The literature on the environmental, social, and governance (ESG) frameworks and CO₂ emissions presents mixed findings across global and country-specific contexts. Refs. [5,28] found that ESG practices contribute to reducing CO₂ emissions globally using bibliometric and ANN methods, while [7] also support ESG’s emission-reducing role within Chinese firms. However, ref. [4], using KRLS and wavelet CQR approaches, found that ESG practices are associated with higher emissions, indicating potential transitional impacts or rebound effects during ESG implementation, especially in the United States and global contexts. These divergent results highlight the need for frequency-based and time-varying approaches to better understand how ESG impact emissions across different regions and policy cycles.

The role of Information and Communication Technology (ICT) in influencing CO₂ emissions is similarly debated within the literature. Studies such as [18,23,29] find that ICT contributes to emission reductions across China, India, and a global sample of 92 countries, emphasizing ICT’s potential in improving energy efficiency and supporting decarbonization. Conversely, refs. [17,19] report that ICT expansion leads to higher CO₂ emissions due to increased energy demand from device usage, data centers, and infrastructure rollouts, particularly within G20 and broader international samples. These contrasting findings underscore the importance of examining the temporal and structural conditions under which ICT can effectively support emission reduction efforts while mitigating rebound effects during digital expansion phases.

Global supply chain (GSC) management consistently shows a positive association with CO₂ emissions in various studies. Refs. [11,24,25,30] demonstrate that GSC activities increase emissions across emerging economies, Japan, the United States, and globally, using methods such as QARDL, SEM, WQQR, and other advanced econometric tools. These findings reflect the carbon-intensive nature of globalized production and logistics systems, which remain difficult to decouple from emissions despite efforts toward greener supply chains. The persistent emissions link in GSC activities indicates the need for policy frameworks that embed low-carbon technologies and ESG principles within supply chain restructuring to align with carbon neutrality goals.

Regarding patent innovations (PAT), the literature also presents mixed evidence on their [31] environmental impacts. Studies by [32,33] found that patent activities reduce CO₂ emissions in the European Union, Arab countries, and globally, suggesting that technological advancements contribute to cleaner production processes. However, refs. [34,35] report that patent activities are associated with increased emissions in BRICS nations and Japan, particularly during periods of rapid industrialization and technology commercialization, reflecting the energy-intensive nature of scaling new technologies. These contrasting results signal a need for deeper examination of the conditions under which patent innovations can support decarbonization, particularly by integrating green patent prioritization within innovation policies to ensure alignment with emissions reduction targets.

Table 1. Summary of past studies.

Author(s)	Period	Nation(s)	Method(s)	Finding(s)
Impact of ESG on CO2
Baratta et al. [5]	Not defined	Global	Bibliometric Analysis	ESG ↓ CO2
Akram et al. [28]	2011–2020	193 countries	ANN	ESG ↓ CO2
Özkan et al. [4]	2002 to 2023	United States	KRLS	ESG ↑ CO2
Qian et al. [7]	Not defined	Chinese companies	Mechanism analysis	ESG ↓ CO2
Impact of ICT on CO2
Zhang & Liu [18]	2000–2010	China	PCSE	ICT ↓ CO2
Añón et al. [17]	1995 to 2010	142 economies	POLS	ICT ↑ CO2
Uddin et al. [19]	1980 to 2019	G20 countries	GMM	ICT ↑ CO2
Özkan et al. [29]	2000–2021	India	MMQR	ICT ↓ CO2
Wang et al. [23]	2006 to 2017	92 countries	Threshold regression	ICT ↓ CO2
Impact of GSC on CO2
Tiwari et al. [25]	1997 to 2020	Emerging economies	QARDL	GSC ↑ CO2
Memari et al. [24]	Undefined	Global	Undefined	GSC ↑ CO2
Maeno et al. [11]	Undefined	Japan	SEM	GSC ↑ CO2
Li & Liu [30]	2000Q1 to 2022Q4.	United States	WQQR	GSC ↑ CO2
Impact of PAT on CO2
Ibrahim et al. [34]	1992 to 2019	BRICS	MMQR	PAT ↑ CO2
Bilgili et al. [32]	1990–2019	European Union	PVAR)	PAT ↓ CO2
Cheng et al. [33]	1990–2019	Arab countries	FMOLS	PAT ↓ CO2
Prakash [31]	1990–2020	190 countries	CCEMG	PAT ↓ CO2
Adebayo & Kirikkaleli [35]	1990–2016	Japan	Wavelet Tools	PAT ↑ CO2

Note: ↑ and ↓ denotes increase and decrease.

presents a summary of past studies.

3. Data and Methods

3.1. Data

In this study, we investigate the co-movement and causality between CO₂ emissions, used as a proxy for environmental sustainability, and Information and Communication Technology (ICT), patent innovations (PAT), global supply chain management (GSC), and Environmental, Social, and Governance (ESG) sustainability. The analysis period spans from 2002Q4 to 2024Q4, with the start date constrained by the availability of ESG data and the end date limited by data availability for ICT and PAT. Except for ICT and PAT, which are converted from their original frequency to quarterly data, all variables are already in quarterly form. Furthermore, CO₂, ICT, PAT, and ESG are transformed using natural logarithms to stabilize variance and interpret elasticities, while GSC remains in its level form. The study begins in 2002Q4 due to the unavailability of ESG data prior to that period, and it ends in 2024Q4 because CO₂ data beyond that point is not available.

Table 2. Data sources.

Abbreviation	Variables	Measurement	Sources
PAT	Patent innovations	Number of patent resident	WDI [36]
ICT	Information and Communication Technology	Individuals using the Internet (% of population)	WDI [36]
CO2	Carbon emissions	Metric tons per capita	OWD [37]
GSC	Global supply chain management	Index
ESG	Environmental sustainability, governance sustainability uncertainty	Index	PU [38]

presents detailed information about the variables under study.

3.2. Empirical Methods

3.2.1. Wavelet Power Spectrum

The wavelet power spectrum ${\left|W_n^x\right|}^2$ captures how the variability of a time series evolves across both time and frequency, revealing periods of high and low volatility within the data. To ensure reliability near the series boundaries, the cone of influence delineates areas where edge effects may distort interpretations. Statistical significance is determined by comparing the observed power against a reference background spectrum $P_f$ under the null hypothesis of stationarity, following the framework outlined by [39].

$$D\left({\displaystyle \frac{{\left|W_n^x\left(s\right)\right|}^2}{{\sigma }_x^2}}<p\right)={\displaystyle \frac{1}{2}}{P}_f{\chi }_v^2,$$

(1)

Here, $P_f$ denotes the average power spectrum, σ represents the variance of the time series, and ${\chi }^2$ refers to the chi-squared distribution used for testing. Significance levels are evaluated using Monte Carlo simulations.

3.2.2. Wavelet Coherency (WTC)

Wavelet coherence is useful for revealing time- and frequency-specific relationships between variables, allowing researchers to detect when and at what scales two series move together or diverge over time [39]. The WTC is defined below:

$$R_n\left(s\right)={\displaystyle \frac{\left|S\left(s^{-1}{W}_n^{xy}\left(s\right)\right)\right|}{\sqrt{S\left(s^{-1}\left|W_n^x\right|\right)S\left(s^{-1}\left|W_n^y\right|\right)}}}$$

(2)

where $S$ represents the process of averaging over both temporal and frequency domains.

3.2.3. Partial Wavelet Coherence (PWC)

Partial wavelet coherence isolates the direct relationship between two variables across time and frequency while controlling for the influence of additional variables [40].

$$R{P}^2\left(y,x_1,x_2\right)={\displaystyle \frac{{\left|R\left(y,x_1\right)-R\left(y,x_2\right)\cdot R{\left(x_2,x_1\right)}^{*}\right|}^2}{{\left[1-R\left(y,x_2\right)\right]}^2{\left[1-R\left(x_2,x_1\right)\right]}^2}},$$

(3)

where $R\left(y,x_1\right) \mathrm{and} R\left(x_2,x_1\right)$ denote wavelet coherence values.

3.2.4. Multiple Wavelet Coherence

Multiple wavelet coherence measures the combined influence of several predictor variables on a target variable across time and frequency, capturing their joint explanatory power over different periods and scales.

$$R{M}^2\left(y,x_1,x_2\right)={\displaystyle \frac{R^2\left(y,x_1\right)+R^2\left(y,x_2\right)-2\mathfrak{R}\left[R\left(y,x_1\right)\cdot R{\left(y,x_2\right)}^{*}\cdot R{\left(x_2,x_1\right)}^{*}\right]}{1-R^2\left(x_2,x_1\right)}}.$$

(4)

3.2.5. Wavelet Granger Causality

Wavelet Granger causality (WGC), as suggested by Adebayo et al. [27], is useful for detecting directional causal relationships between variables across different time scales, revealing how drivers impact outcomes over short-, medium-, and long-term horizons. This approach helps policymakers and researchers identify frequency-specific causality patterns, supporting targeted interventions aligned with dynamic system behaviors. Adebayo et al. [27] defined WGC as follows:

$$W_Y^s(t)={\alpha }^s+\sum _{i=1}^p{\beta }_i^s{W}_Y^s(t-i)+\sum _{i=1}^p{\gamma }_i^s{W}_X^s(t-i)+{ϵ}_t^s$$

(5)

Here, $W_Y^s\left(t\right) \mathrm{and} W_X^s\left(t\right)$ refer to the wavelet coefficients corresponding to scale s and time point t. The terms ${\alpha }^s,{\beta }_i^s,{\gamma }_i^s$ represent the coefficients linked to the wavelet-transformed components, while ${ϵ}_t^s$ denotes the residual error at the specified scale.

4. Findings and Discussion

4.1. Descriptive Statistics

Table 3. Descriptive statistics.

	CO2	ICT	PAT	ESG	GSC
Minimum	1.235	1.664	3.609	2.377	−1.313
Maximum	2.128	4.356	7.102	3.894	4.251
1. Quartile	1.721	3.082	4.967	2.765	−0.458
3. Quartile	2.016	4.149	6.852	3.283	0.298
Mean	1.854	3.518	5.877	3.033	0.127
Median	1.957	3.846	6.311	2.950	−0.163
SE mean	0.025	0.087	0.118	0.037	0.109
LCL mean	1.804	3.344	5.642	2.959	−0.088
UCL mean	1.904	3.691	6.112	3.107	0.343
Variance	0.056	0.679	1.244	0.123	1.048
Stdev	0.238	0.824	1.116	0.351	1.024
Skewness	−1.044	−0.927	−0.584	0.433	1.950
Kurtosis	−0.070	−0.536	−1.121	−0.775	3.856
JB	16.712 ***	14.032 ***	9.549 ***	4.823 *	118.04 ***

Note: *** p < 1% and * p < 10%.

presents the descriptive statistics for the variables used in this study, namely, CO₂ emissions, ICT, PAT, ESG, and GSC. Starting with CO₂, the variable shows a relatively low dispersion (standard deviation of 0.238) and a negatively skewed distribution (skewness = −1.044), indicating a longer left tail and concentration of higher values. The Jarque–Bera (JB) statistic (16.712) confirms a statistically significant deviation from normality at the 1% level, suggesting non-normality in emissions data—likely influenced by structural shifts in energy usage or environmental policy changes over time. ICT also displays negative skewness (−0.927), indicating that the majority of the observations lie above the mean. It has a relatively higher variance (0.679) and standard deviation (0.824), suggesting moderate fluctuation in digital infrastructure metrics across time or regions. The non-normality is statistically supported by a JB value of 14.032. This aligns with the known uneven digital penetration across economies and periods. PAT exhibits the highest mean (5.877) and median (6.311), implying an overall higher level of innovation output. Its skewness is also negative (−0.584), but less pronounced than CO₂ and ICT, with kurtosis indicating a platykurtic (flat-tailed) distribution (−1.121). The JB test (9.549) still suggests non-normality at the 1% significance level, possibly reflecting bursts of innovation during certain periods followed by stagnation. In contrast, ESG shows a mildly positive skew (0.433) and slightly lower variability (standard deviation of 0.351), indicating a relatively stable distribution around the mean (3.033). Its JB statistic (4.823) is significant at the 10% level, suggesting only marginal departure from normality, possibly due to standardized ESG scoring systems or regulatory convergence across regions. The most distinct behavior is observed in GSC, with a strong positive skew (1.950) and high kurtosis (3.856), implying a highly right-skewed and leptokurtic distribution. This suggests infrequent but extreme values—potentially linked to pandemic-driven supply chain shocks or geopolitical disruptions. The JB statistic of 118.04 strongly confirms non-normality. Notably, its mean (0.127) and median (−0.163) are close to zero, implying that while most GSC values are near average, extreme supply chain fluctuations skew the overall distribution.

4.2. Nonlinearity Test Result

Table 4. BDS test result.

Dimensions	CO2	ICT	PAT	ESG	GSC
M2	24.584 ***	24.318 ***	36.582 ***	19.929 ***	9.8861 ***
M3	26.176 ***	25.852 ***	39.070 ***	19.716 ***	10.312 ***
M4	28.181 ***	27.744 ***	42.225 ***	19.922 ***	10.494 ***
M5	31.132 ***	30.504 ***	46.860 ***	20.601 ***	10.908 ***
M6	35.196 ***	34.294 ***	53.253 ***	21.851 ***	11.368 ***

Note: *** p < 1%.

reports the results of the Brock–Dechert–Scheinkman (BDS) test across embedding dimensions M2 to M6 for CO₂, ICT, PAT, ESG, and GSC. The consistently significant results at the 1% level across all dimensions indicate the rejection of the null hypothesis of independent and identically distributed (i.i.d.) residuals, suggesting the presence of nonlinearity and complex dependence structures in all the series examined. Notably, PAT exhibits the highest BDS statistics across dimensions, implying stronger nonlinear dynamics compared to the other variables, while GSC shows the lowest yet significant values, indicating milder but present nonlinear dependencies. These findings justify the application of nonlinear and nonparametric methods, such as wavelet analyses.

4.3. Stationarity Test Result

Figure 1 illustrates the wavelet Zivot–Andrews. Unlike the conventional ZA, this test captures stationarity across different scales. The results are shown across MODWT detail levels (D1–D5) for CO₂, ESG, GSC, ICT, and PAT. Each panel shows the test statistic (vertical axis) across decomposition levels (horizontal axis), with break dates labeled on the bars, indicating the most significant structural breaks in each frequency band. The red dotted, blue dashed, and black solid lines represent the 1%, 5%, and 10% critical values. Values below these lines indicate rejection of the null hypothesis of a unit root with structural breaks, suggesting stationarity in the corresponding frequency band and variable. Notably, CO₂, ICT, and PAT display significant statistics across D1 and D2, indicating that at high and medium-high frequencies, these series are stationary with structural breaks, while ESG and GSC generally exhibit higher test statistics, suggesting persistence in shocks across some frequency bands. The identified break dates, such as 2005-03 for PAT and ICT and 2023-09 for CO₂ and GSC, align with global and national economic or environmental events, illustrating the impact of structural shifts on these variables’ dynamics across time–frequency domains.

4.4. Wavelet Power Result

In this study, we check the volatility of the variables using the wavelet power spectrum (WPS). Figure 2 presents the results of the WPS. For CO₂ (see Figure 2a), the result shows high power in the short-term frequency bands (periods 4–8) before 2010, indicating significant short-term volatility in emissions during China’s rapid industrialization and urbanization phases, which saw heavy coal use and energy consumption spike. After 2012, the spectrum shows predominantly low power across all frequencies, reflecting the effects of China’s environmental regulations, industrial upgrading, and energy efficiency measures stabilizing CO₂ emissions growth. ICT (see Figure 2b) similarly exhibits short-term high power before 2010, reflecting volatility linked to China’s aggressive expansion in broadband and mobile infrastructure, rapid digital adoption, and global ICT market fluctuations. The reduced power post-2012 suggests the maturation of ICT penetration and policy stabilization within China, leading to smoother growth patterns in the sector.

For PAT (see Figure 2c), the spectrum shows substantial high power pre-2010, aligned with China’s push for technological self-reliance and its entry into the World Trade Organization, which spurred patent filings and R&D volatility during the catch-up phase. The muted power post-2012 indicates a more stable innovation environment, reflecting consistent policy support for R&D and the emergence of high-tech zones contributing to steady patent output. For ESG (see Figure 2d), we observed episodic volatility across multiple frequencies between 2005 and 2012, linked to China’s initial phases of environmental governance, introduction of ESG-related frameworks, and efforts to align corporate practices with sustainability targets. The presence of scattered power patches after 2012 also reflects intermittent ESG policy adjustments and enforcement challenges in different provinces as China ramped up its green finance and carbon neutrality goals. Likewise, GSC (see Figure 2e) shows significant high power around 2008–2010 and post-2018 across medium- to long-term frequencies (periods 8–16), indicating substantial volatility in China’s export-driven supply chains during the global financial crisis and subsequent recovery. The heightened power post-2018 corresponds to US–China trade tensions, COVID-19 pandemic disruptions, and global logistics bottlenecks impacting China’s role in the global supply chain, consistent with findings on supply chain fragility under external shocks.

4.5. Wavelet Coherence Result

The study also proceeds to examine the co-movement and lead/lag association between CO₂ and its determinants (ESG, GSC, PAT, and ICT) in China (see Figure 3).

Figure 3a shows the wavelet coherence (WTC) analysis between CO₂ emissions and global supply chain (GSC) management in China using quarterly data from 2002Q4 to 2024Q4. Notably, during the period around 2010–2013, significant coherence is observed within the eight-period band (approximately eight quarters or 2 years), indicated by the black contours within the cone of influence. The upward-pointing arrows in this region suggest in-phase co-movement, implying that increases in CO₂ emissions were accompanied by increases in GSC activities during this period, potentially reflecting China’s aggressive manufacturing expansion and export-led strategies following the 2008 global financial crisis [11,14]. Outside this band, coherence is relatively weak, indicating that the relationship between CO₂ emissions and GSC management was less synchronized during other periods, aligning with findings that the carbon intensity of supply chains fluctuates based on trade policies and industrial upgrading cycles [12]. From 2018 to 2023, significant coherence emerges around the 8–16 period band (approximately 8–16 quarters or 2–4 years), particularly near the boundary of the cone of influence, suggesting emerging long-term co-movement likely driven by China’s decarbonization policies and global supply chain restructuring [11]. The arrows in these regions point predominantly rightward and downward, indicating a positive, in-phase relationship with a lead–lag structure, where changes in GSC management precede adjustments in CO₂ emissions with some lag. This pattern may reflect the time required for supply chain reconfigurations, cleaner logistics, and green procurement practices to translate into measurable environmental impacts, consistent with the evolving nature of supply chain decarbonization in China.

Figure 3b shows the WTC between CO₂ emissions and environmental, social, and governance (ESG) sustainability indicators in China using quarterly data. During 2007–2016, significant coherence is observed within the 4–8 period band (approximately 1–2 years, or 4–8 quarters), where arrows predominantly point leftward and downward, indicating a negative, out-of-phase relationship with ESG changes leading CO₂ emissions. This suggests that improvements in ESG sustainability practices were associated with subsequent reductions in CO₂ emissions, consistent with the effectiveness of China’s ESG-driven environmental regulations and corporate sustainability actions during this period [4,8]. The downward component implies a lead–lag effect, where ESG initiatives influenced emissions reductions after a short adjustment period as industries adapted practices under strengthened environmental governance during China’s 11th and 12th Five-Year Plans [41]. In contrast, during 2019–2024, the analysis reveals strong coherence within the 8–16 period band (approximately 2–4 years, or 8–16 quarters), with arrows predominantly pointing rightward and downward, indicating a positive, in-phase relationship with ESG changes leading CO₂ emissions. This suggests that increases in ESG sustainability metrics coincided with increases in CO₂ emissions in the medium term, reflecting a period where intensified ESG activities, including infrastructure investments for green transitions and ESG-compliant industrial expansions, may have led to temporary rises in emissions due to construction activities and industrial upgrading [42].

Figure 3c shows the WTC between CO₂ emissions and ICT development in China using quarterly data. From 2006 to 2016, significant coherence is observed across the 8–16 period band (approximately 2–4 years, or 8–16 quarters), with arrows predominantly pointing rightward and downward, indicating a positive, in-phase relationship where ICT developments lead CO₂ emissions. This suggests that the expansion of ICT infrastructure, data centers, and device manufacturing during China’s rapid digitalization phase contributed to increased energy consumption and CO₂ emissions despite the sector’s potential for efficiency improvements [18]. The downward component of the arrows indicates a lead–lag structure, implying that ICT investments and expansions often preceded rises in emissions with a time lag, reflecting the energy-intensive nature of ICT rollouts during China’s industrial transformation and efforts to reduce the urban–rural digital divide under the 12th Five-Year Plan. In the period 2018 to 2024, the coherence becomes prominent within the 4–8 period band (approximately 1–2 years, or 4–8 quarters), with rightward-pointing arrows indicating a continued positive, in-phase relationship between ICT growth and CO₂ emissions. This indicates that ongoing ICT development, including 5G network deployment, smart city initiatives, and IoT expansion, is closely associated with increases in CO₂ emissions, highlighting the rebound effects of digitalization on energy demand [43,44].

Figure 3d shows the WTC between CO₂ emissions and patent innovations (PAT) in China over the study period using quarterly data. Between 2006 and 2016, significant coherence is observed within the 8–16 period band (approximately 2–4 years, or 8–16 quarters), where arrows predominantly point rightward and downward, indicating a positive, in-phase relationship with patent innovations leading CO₂ emissions. This suggests that increased patenting activity, including green and non-green technologies, was associated with rising CO₂ emissions, reflecting China’s rapid industrialization and innovation-driven growth strategy that heavily relied on energy-intensive manufacturing during this period [33]. The downward arrow component indicates a lead–lag structure, implying that the surge in innovation activity often preceded increases in emissions as patented technologies were commercialized and scaled in industries. Furthermore, the study shows that from 2018 to 2024, the plot shows continued significant coherence within the 4–8 period band (approximately 1–2 years, or 4–8 quarters), with arrows still pointing rightward, indicating a continued positive, in-phase association between patent innovations and CO₂ emissions. This pattern suggests that while patenting activity, including in clean energy technologies, has expanded in China, it has been accompanied by increases in CO₂ emissions, potentially due to the energy-intensive processes involved in the production and commercialization of patented technologies and the rebound effects of technological advancement [45,46].

4.6. Multiple Wavelet Coherence

The study also used the multiple wavelet coherence to capture the effect of X1 on Y while considering the role of X2. Figure 4 presents the MWC results.

Figure 4a (MWC: CO₂-GSC-ESG) examines the impact of global supply chain (GSC) management on CO₂ emissions while considering the role of ESG sustainability in China using quarterly data from 2002Q4 to 2024Q4. Significant coherence within the 4–8 period band (1–2 years) and 8–16 period band (2–4 years), particularly during 2008–2015 and 2018–2024, indicates that GSC activities have a significant impact on CO₂ emissions during periods of industrial expansion and trade growth. The presence of ESG as a moderating factor is evident during these periods, reflecting the partial effectiveness of ESG frameworks in China’s supply chain-related emission management. Figure 4b (MWC: CO₂-ICT-ESG) and Figure 4c (MWC: CO₂-ICT-GSC) illustrate the impact of ICT development on CO₂ emissions while considering the role of ESG sustainability. Between 2008 and 2015, significant coherence within the 8–16 period band highlights a strong impact of ICT expansion, including data centers and infrastructure rollouts, on emissions, with ESG sustainability helping to moderate these effects in some periods. From 2018 to 2024, coherence within the 4–8 period band indicates that ICT development continues to have a notable impact on emissions in the medium term, with ESG practices and GSC management influencing this relationship and emphasizing the need to align digitalization with low-carbon transitions to avoid rebound effects.

Figure 4d (MWC: CO₂-PAT-ESG) and Figure 4e (MWC: CO₂-PAT-GSC) analyze the impact of patent innovations on CO₂ emissions while considering the role of ESG sustainability and GSC management. During 2008–2015, coherence within the 8–16 period band suggests that increases in patent activity, including green technologies, had a significant impact on emissions as innovations were commercialized within energy-intensive industries. From 2018 to 2024, coherence within the 4–8 period band shows that patent-driven innovation continues to have an impact on emissions, with ESG sustainability and GSC management providing partial mitigation but not fully decoupling innovation from emissions trends. Figure 4f (MWC: CO₂-ESG-GSC) complements these insights by illustrating the impact of ESG sustainability on CO₂ emissions while considering the role of GSC management, emphasizing that ESG efforts can influence emissions but remain highly dependent on concurrent supply chain restructuring to effectively support China’s carbon neutrality targets.

4.7. Partial Wavelet Coherence

The study also used the partial wavelet coherence (PWC) to capture the effect of X1 on Y while neglecting the role of X2. Figure 5 presents the PWC results.

Figure 5a (PWC: CO₂-ICT-PAT) shows the impact of ICT development on CO₂ emissions while neglecting the role of patent innovations in China using quarterly data from 2002Q4 to 2024Q4. During 2008–2015, limited but notable coherence within the 4–8 quarter band (approximately 1–2 years) indicates significant impact, likely reflecting China’s rapid ICT infrastructure expansion aligning with changes in energy demand and industrial activities during the digital economy transition. Figure 5b (PWC: CO₂-ICT-ESG) shows the impact of ICT development on CO₂ emissions while neglecting the role of ESG sustainability in China using quarterly data. From 2007 to 2017, strong coherence within the 8–16 quarter band (approximately 2–4 years) indicates significant impact, likely driven by the rapid scale-up of data centers and ICT services increasing electricity demand in the absence of ESG moderation during industrial growth phases.

Figure 5c (PWC: CO₂-ICT-GSC) shows the impact of ICT development on CO₂ emissions while neglecting the role of global supply chain (GSC) management in China using quarterly data. During 2018–2024, coherence within the 4–8 quarter band (1–2 years) indicates significant impact, possibly reflecting the expansion of smart city initiatives and 5G infrastructure influencing energy consumption patterns. Figure 5d (PWC: CO₂-PAT-ICT) shows the impact of patent innovations on CO₂ emissions while neglecting the role of ICT development in China using quarterly data. Figure 5e (PWC: CO₂-PAT-ESG) shows the impact of patent innovations on CO₂ emissions while neglecting the role of ESG sustainability in China using quarterly data. Between 2008 and 2015, coherence within the 8–16 quarter band (2–4 years) in both figures indicates significant impact, likely reflecting periods when patent-driven technologies were commercialized in energy-intensive sectors during China’s industrial upgrading.

Figure 5f–i (PWC: CO₂-PAT-GSC, CO₂-ESG-ICT, CO₂-ESG-PAT, CO₂-ESG-GSC) show the impact of the primary variable (PAT, ESG, or ICT) on CO₂ emissions while neglecting the roles of GSC, ICT, and ESG in various combinations in China using quarterly data. Significant impact within the 4–8 (1–2 years) and 8–16 (2–4 years) quarter bands is observed across these combinations, highlighting the persistent influence of technological developments, supply chain structures, and sustainability practices on emission dynamics during China’s transition toward a low-carbon economy. Figure 5j (PWC: CO₂-GSC-ICT) shows the impact of global supply chain (GSC) management on CO₂ emissions while neglecting the role of ICT development in China using quarterly data. Notable coherence around 2010–2013 within the 8–16 quarter band (2–4 years) indicates significant impact, likely tied to post-crisis export-led recovery and industrial expansion. Figure 5k (PWC: CO₂-GSC-PAT) shows the impact of GSC management on CO₂ emissions while neglecting the role of patent innovations, with minor signals around 2012–2015 indicating limited but notable impact, possibly linked to adjustments in manufacturing supply chains. Figure 5l (PWC: CO₂-GSC-ESG) shows the impact of GSC management on CO₂ emissions while neglecting the role of ESG sustainability, with faint signals around 2008–2010 within the 4–8 quarter band indicating minimal but identifiable impact, reflecting the limited effectiveness of GSC management in influencing emissions without ESG frameworks in China.

4.8. Wavelet Granger Causality Result

Next, we examine the causal association among the variables using the recently developed WGC by [27] (see Figure 6 and

Table 5. Wavelet Granger causality result.

Frequency	Dependent Variables	Independent Variables
		CO2	ESG	GSC	PAT	ICT
Original level	CO2		0.7456	0.7172	0.0320	0.0353 **
	ESG	0.530		0.1520	0.3819	0.1810
	GSC	0.214	0.4000		0.5024	0.9890
	PAT	0.952	0.2930	0.6365		0.6737
	ICT	0.189	0.7557	0.6668	0.0213
Level 1	CO2		0.001	0.590	0.241	0.402
	ESG	0.174		0.392	0.129	0.104
	GSC	0.110	0.365		0.136	0.127
	PAT	0.293	0.001 ***	0.555		0.950
	ICT	0.441	0.001 ***	0.504	0.954
Level 2	CO2		0.008	0.765	0.742	0.760
	ESG	0.388		0.204	0.385	0.386
	GSC	0.227	0.719		0.126	0.146
	PAT	0.648	0.010 **	0.936		0.527
	ICT	0.978	0.013 **	0.915	0.680
Level 3	CO2		0.170	0.000 ***	0.000 ***	0.000 ***
	ESG	0.003 ***		0.011 **	0.002 ***	0.000 ***
	GSC	0.000 ***	0.366		0.000 ***	0.000 ***
	PAT	0.001 ***	0.052 *	0.000 ***		0.001 ***
	ICT	0.000 ***	0.070 *	0.000 ***	0.002 ***
Level 4	CO2		0.000 ***	0.000 ***	0.000 ***	0.000 ***
	ESG	0.000 ***		0.000 ***	0.000 ***	0.000 ***
	GSC	0.000 ***	0.000 ***		0.000 ***	0.000 ***
	PAT	0.000 ***	0.000 ***	0.000 ***		0.006 ***
	ICT	0.000 ***	0.000 ***	0.000 ***	0.002 ***
Level 5	CO2		0.000 ***	0.000 ***	0.001 ***	0.000 ***
	ESG	0.000 ***		0.000 ***	0.000 ***	0.000 ***
	GSC	0.000 ***	0.001 ***		0.000 ***	0.000 ***
	PAT	0.000 ***	0.000 ***	0.000 ***		0.002 ***
	ICT	0.000 ***	0.000 ***	0.000 ***	0.051 *

Note: Significance levels are indicated as *** p < 0.01, ** p < 0.05, and * p < 0.10, marking the rejection of the null hypothesis of no causality. Level 1 and Level 2 capture short-term dynamics, Level 3 represents medium-term movements, while Level 4 and Level 5 correspond to long-term relationships.

). This approach allows for the examination of causal relationships across different time scales, capturing both short-term and long-term dynamics. This approach is particularly useful for non-stationary data and complex systems where traditional Granger causality may fail to detect time-varying interactions.

At Level 1 and Level 2 (capturing short-term dynamics: 4–8 quarters, or 1–2 years), there are a few scattered significant relationships. For example, ESG Granger-causes PAT and ICT at the 1% level at Level 1, and at Level 2, ESG significantly Granger-causes PAT and ICT at the 5% level. These findings suggest that short-term fluctuations in ESG sustainability influence patenting activity and ICT developments, reflecting ESG’s role in shaping technological directions within China’s short-term policy cycles. However, the influence of these variables on CO₂ emissions in the short term remains generally weak, indicating that emission responses to these drivers require longer adjustment periods to manifest significantly. In contrast, Levels 3, 4, and 5 (capturing medium- to long-term dynamics: 8–32 quarters, or 2–8 years) reveal pervasive and robust causal relationships across the variables, with widespread significance at the 1% level. Notably, CO₂ emissions are significantly Granger-caused by ESG, GSC, PAT, and ICT in the medium- and long-term frequencies, indicating that sustainability initiatives, supply chain activities, technological innovation, and digitalization collectively drive emissions over extended horizons in China. Similarly, ESG, GSC, PAT, and ICT are all significantly influenced by each other in the long term, reflecting the deep interconnectedness between environmental governance, technological advancement, supply chain restructuring, and emission reduction pathways in China. These results underscore the importance of adopting a frequency-specific lens when designing policies, emphasizing that while short-term impacts on emissions may be limited, medium- and long-term synergies between ESG, ICT, patenting, and supply chain management are critical for shaping China’s low-carbon transition and ensuring that economic and technological advancements align with carbon neutrality goals.

5. Conclusions and Policy Recommendations

5.1. Conclusions

Harnessing ICT and global supply chain (GSC) advancements without aligning them with robust ESG strategies risks locking China into a trajectory of elevated CO₂ emissions. This highlights the urgent need to integrate digitalization and supply chain expansion with sustainability objectives to ensure a genuinely low-carbon future. To investigate this, we employed a suite of wavelet-based tools to analyze the drivers of CO₂ emissions in China using quarterly data from 2002Q4 to 2024Q4. The wavelet coherence analysis reveals that GSC management, ICT development, and patent innovations exhibit consistent, positive, and in-phase relationships with CO₂ emissions—particularly during the periods 2008–2016 and 2018–2024—indicating their contribution to rising emissions. In contrast, ESG improvements were associated with reduced emissions between 2007 and 2016 but shifted to a positive association after 2018, likely due to transitional implementation effects. Multiple wavelet coherence analysis confirms that although ESG and GSC management partially mitigate emissions, they have not succeeded in fully decoupling economic growth from environmental degradation during these intervals. Partial wavelet coherence (PWC) results further show that ICT, patent innovations, ESG, and GSC significantly influence CO₂ emissions within the 4–8 quarter (1–2 years) and 8–16 quarter (2–4 years) bands. Finally, wavelet Granger causality analysis indicates that ESG exerts a weak short-term causal effect on ICT and patent activity, as well as on CO₂ emissions, while medium- to long-term causal linkages among ESG, GSC, PAT, ICT, and CO₂ are found to be robust.

5.2. Policy Recommendations

Given that global supply chain management shows a positive, in-phase relationship with CO₂ emissions during China’s post-crisis industrial expansion (2010–2013) and again during 2018–2023, policymakers should strengthen green supply chain frameworks under China’s Green Supply Chain Pilot Programs and the Dual Carbon strategy. This includes mandating carbon accounting across domestic and cross-border supply chains, incentivizing green logistics and procurement practices, and ensuring that supply chain restructuring under industrial upgrading programs integrates emission reduction goals. Additionally, aligning trade and export promotion policies with carbon neutrality targets will help decouple GSC expansion from emissions growth, particularly by prioritizing low-carbon manufacturing and cleaner logistics pathways in China’s Belt and Road Initiative corridors.

The result on ESG indicate that while ESG improvements reduced emissions during 2007–2016, post-2018 ESG activities have coincided with rising emissions due to transition-related construction and infrastructure expansion. This underscores the need to enhance the effectiveness of ESG frameworks in driving decarbonization outcomes. China should move ESG from reporting compliance toward enforceable standards tied to emissions reductions, integrating carbon intensity and lifecycle emissions metrics within ESG disclosure requirements currently expanding under the China Securities Regulatory Commission’s ESG frameworks. Additionally, green finance policies should link ESG performance to access to preferential credit and subsidies, ensuring that infrastructure and industrial expansions under ESG umbrellas contribute effectively to emissions reductions rather than rebound effects [47].

The analysis showing a persistent positive, in-phase relationship between ICT development and CO₂ emissions highlights the rebound effects of digitalization on energy demand. To address this, China should accelerate green ICT policies by mandating energy efficiency standards and renewable energy sourcing in 5G deployment, data center operations, and smart city infrastructure under the “Digital China” strategy and the 14th Five-Year Plan. Establishing carbon neutrality pathways for the ICT sector, expanding pilots for low-carbon data centers, and providing incentives for ICT firms adopting clean energy and energy-efficient practices will help mitigate emissions while maintaining the momentum of digital transformation critical to economic modernization.

Finally, the finding that patent innovations are positively associated with CO₂ emissions—even in the case of green technologies—underscores the energy-intensive nature of innovation commercialization in China. This highlights a critical tension between technological advancement and environmental sustainability. To ensure that innovation aligns with China’s “dual carbon” targets—peaking carbon emissions before 2030 and achieving carbon neutrality by 2060—policymakers should integrate decarbonization goals into the country’s intellectual property and industrial strategies. Specifically, the China National Intellectual Property Administration (CNIPA) should further expand fast-track channels for green patents and establish stronger linkages between patent commercialization and participation in the national carbon market. Additionally, under strategic frameworks such as Made in China 2025 and the National Innovation-Driven Development Strategy, firms scaling up patented technologies should be mandated to implement emissions offset mechanisms or adopt clean production standards. These measures are essential to ensure that innovation-driven industrial upgrading supports rather than undermines the country’s long-term low-carbon transition, mitigating the risk of technology-induced rebound emissions.

5.3. Limitation and Future Directions

A key limitation of this study is that while wavelet tools effectively capture time–frequency dynamics between CO₂ emissions and drivers such as ICT, ESG, and GSC, they do not fully account for sector-specific differences, potential nonlinear feedback effects, or structural policy shifts within China that may influence these relationships. Additionally, the analysis identifies significant coherence and causality but does not quantify the exact magnitude of each driver’s contribution to emissions, which limits direct policy targeting. Future research should incorporate wavelet-based quantile or nonlinear frameworks to explore heterogeneous impacts across regions and sectors, while integrating structural break tests and machine learning approaches to capture complex dynamics. Expanding the scope to include financial development, renewable energy transitions, and regional policy interactions will further support designing effective low-carbon strategies aligned with China’s carbon neutrality goals.