Impact of Digitalization on Carbon Emissions in Guangdong’s Manufacturing Sector: An Input–Output Perspective

Abstract

As global pressure to reduce emissions intensifies, China is increasingly turning to digital technologies to drive sustainable industrial development, aiming to boost production while keeping carbon emissions in check. This study takes a micro-level approach by dividing the industry into 17 sectors and applying an environmentally-extended input–output (EEIO) model combined with structural decomposition analysis (SDA) to quantify the impact of digital transformation on carbon emissions across sectors. This study used input–output data from 2012 and 2017. The results indicate that (1) technological improvements driven by digitalization play a key role in reducing industrial carbon emissions, and (2) while high-carbon sectors show substantial emission reductions due to digital transformation, industries such as textiles—where digital adoption is more challenging—exhibit only limited improvements. These findings underscore the need to further advance technological upgrading and transformation in less digitally integrated sectors.

1. Introduction

As a key pillar of the global economy, the manufacturing industry plays a crucial role in driving economic growth and generating more employment. However, it is a major contributor to the global greenhouse gas (GHG) emissions [1]. The production and transportation of goods within the manufacturing industry are all dependent on fossil fuels, which will produce lots of GHGs such as CO₂. According to the IEA’s (International Energy Agency) report, carbon emissions from manufacturing now account for more than 30% of the global greenhouse gas emissions, especially in high energy consumption sector, such as steel, chemical, and building materials manufacturing [2].

Amid growing concerns over global warming and climate change, along with the continuing progress of International Environmental Agreements, reducing carbon emissions has become an international consensus. Under this background, digital technology is available that can control the increase in carbon emissions while promoting economic growth (Danish et al., 2018) [3]. At the same time, with the advent of the Industry 4.0 era, digital technologies are increasingly being integrated into both everyday life and industrial production, including within the manufacturing sector. Industrial digitalization is widely recognized as a key method for addressing the dual challenge of enhancing productivity and reducing emissions [4]. Industrial digitalization can not only optimize the production process and improve the efficiency of energy use but can also reduce the waste of resources and achieve the goal of carbon dioxide emission reduction through the introduction of the Internet of Things (IoT), big data analysis, and other technological means in the manufacturing industry in the areas of product design, production, and transportation [5].

As the center of global manufacturing, China faces substantial pressure to reduce carbon emissions while sustaining economic growth. From 2002 to 2019, China’s carbon emissions experienced a long-term growth; although the growth rate slowed down in 2020, the country still accounted for approximately 30% of global carbon emissions [6]. To lower carbon emissions and foster sustainable development, China has set ambitious national targets: to achieve the carbon peak by 2030 and carbon neutrality in 2060. These challenging targets call for extensive reforms and enhancements in both the industrial and energy sectors, especially within manufacturing [7]. The “Made in China 2025” strategy also clearly emphasizes deeply integrating information technology with manufacturing, advancing intelligent and green manufacturing practices, and accelerating the transformation and upgrading of traditional industries [8]. In addition, the subsequent “New Quality Productive Forces” advocates for the development of a “Digital China” highlighting the role of digitalization and technology innovation in fostering green productivity.

Guided by these national strategies, manufacturing enterprises in China are increasingly embracing digital technologies to pursue low-carbon development while enhancing productivity. Especially in energy-intensive industries, digital technologies have been widely applied in energy-intensive sectors to optimize energy consumption and reduce emissions. However, due to the different structures of enterprises and industries sectors, there are significant differences in the performance and effectiveness of digital transformation, which is a phenomenon requiring in-depth study.

Guangdong is a major manufacturing province located in the southern coastal region of China (Figure 1), accounting for nearly 4% of the world’s manufacturing industry. The province is home to Shenzhen, the hub of China’s reform and opening up, which is not only a cluster of high-tech industries but also a frontline of manufacturing innovation and development. In response to national calls for “smart manufacturing” and “green development”, Guangdong’s manufacturing sector has actively embraced industrial digital transformation. Also as the Figure 2 and

Table A1. Carbon emission intensity of Guangdong Province, 2012–2022.

Year	Carbon Emissions (million Tons CO2)	GDP (Billion CNY)	Carbon Emission Intensity (Tons/ Billion CNY)
2012	62,253.41	57,007.7	1.092018
2013	62,637.52	62,503.4	1.002146
2014	63,078.54	68,173	0.925272
2015	63,485.74	74,732.4	0.849508
2016	65,295.89	82,163.2	0.79471
2017	68,253.42	91,648.7	0.744729
2018	70,549.51	99,945.2	0.705882
2019	69,855.48	107,987	0.646888
2020	70,717.62	111,152	0.636224

illustrates, carbon emission intensity in Guangdong Province has been steadily declining. By integrating intelligent equipment and digital management systems, many industries have succeeded in improving production efficiency and achieving carbon emissions reduction. In 2023, the value added of Guangdong’s manufacturing industry will reach 443.979 billion, representing 32.7% of the province’s GDP—the second highest in China—and accounting for 13.5% of the national total [9]. However, there are still differences in the breadth and depth of the application of digital technology in different industries and enterprises, which leads to different carbon emission reduction effects. As one of the pioneering regions of the “dual-carbon” goal, the study of the carbon emission characteristics of Guangdong manufacturing enterprises is of great significance as a reference for low-carbon manufacturing in other regions of China. Guangdong also has an advantage in promoting industrial digitalization. Ding et al. highlight that the province leverages its highly skilled pool and high-tech foundation to promote industrial modernization through digital means [10]. Meanwhile, Nie et al. found that during the promotion of China’s digital local policies, the carbon emissions in Guangdong Province have been particularly significantly reduced [11]. From this perspective, Guangdong represents a highly appropriate and representative case study for investigating the relationship between industrial digitalization and carbon emission reduction.

Although previous studies have examined the emission reduction effects of digitalization, most have focused on the macro layer or a single industry, lacking a comparative analysis across different industry sectors. Taking listed companies in the manufacturing industry of Guangzhou as the research object, this study uses the carbon emissions data of different sectors in the manufacturing industry to measure the degree of digitalization of enterprises through the digital input ratio and further adopts the fixed effect model to analyze the extent of the impact of industrial digitalization on the reduction in carbon emissions, as well as the difference in the impact of carbon emissions on different sectors of the manufacturing industry. Through the empirical study of Guangzhou’s manufacturing industry, this thesis aims to provide policy recommendations for the digital manufacturing transformation in Shenzhen and other regions and to contribute data support as well as a theoretical basis for the realization of the national “dual-carbon” goal.

2. Literature Review

2.1. Industrial Digitalization and Industrial Transformation

According to the 2017 White Paper on China’s Digital Economy, the digital economy consists of two main components: digital industrialization and industrial digitalization [12]. Digital industrialization refers to the development of the information industry, primarily including electronic information manufacturing and software services. In contrast, industrial digitalization refers to the enhancement of output and efficiency through the integration of digital technologies into traditional industries. This process not only improves production quantity and efficiency but also significantly enhances the value-added content of industrial activities. This study focuses on the industrial digitalization aspect, which is widely regarded as an important means of transforming the manufacturing sector by converting traditional industries into cleaner, more advanced, and environmentally friendly high-tech industries.

In China, the transformation and upgrading of the manufacturing industry have been consistently emphasized as a focus of national economic development. The Made in China 2025 strategy highlights the importance of integrating information technology and promoting industrial upgrading around key sectors of manufacturing. The digitalization of Chinese enterprises often starts from the production side, which is mainly manifested in the data and visualization of the production process, the installation of detection devices for the production equipment of enterprises, and the monitoring of their energy consumption and emissions, so as to improve the production efficiency and better manage the energy use in the production of enterprises. Li Tianren [13] argues that the integration of digital technology with industry changes enterprise production models and advances the sharing of information and the exchange of technical expertise, thereby driving the green transition and sustainable growth of traditional industries.

2.2. Influence Mechanism of Digital Technology on Carbon Emissions in Manufacturing Industry

Empirical analysis shows that digital technology can reduce specific energy consumption and achieves emission reduction effects by improving production efficiency, optimizing resource allocation, and promoting green innovation. Especially in high-energy-consuming industries, smart manufacturing technology has been proven to have greater potential for reducing carbon emissions [13]. Taking the power industry as an example, its digital transformation is primarily manifested in the construction of smart grids and the deployment of digital equipment. On the one hand, the application of intelligent dispatch and real-time monitoring technologies has improved the operational efficiency of power systems and reduced energy waste; on the other hand, digital control systems have also encouraged enterprises to replace fossil fuels with clean energy in production processes, thereby reducing the total CO₂ emissions associated with the power industry from the source [14].

Meanwhile, the application of digital technologies in manufacturing spans multiple stages, including production, logistics, and sales [15]. For example, quality monitoring reduces raw material waste, intelligent warehousing optimizes logistics routes and improves inventory turnover efficiency, while data analysis at the sales end helps accurately analyze the market and achieve supply–demand balance. The systematic optimization of these multiple stages collectively achieves improved resource utilization efficiency and reduced indirect carbon emissions.

However, the emissions reduction effectiveness of digitalization varies across industries. This disparity stems from inherent differences in production methods, technological foundations, and digital infrastructure across sectors. Zhang et al. [16] found that high-tech industries, due to their higher digital maturity and complex production processes, are more likely to achieve significant carbon emissions reductions through digital transformation. In contrast, labor-intensive industries and low-tech industries have relatively limited emission reduction effects. These findings suggest that when assessing the impact of digitalization on carbon emissions reduction, it is important to emphasize industry-level differences. If estimates are made solely from a macro perspective, structural differences between industries in terms of digitalization-driven pathways and emission reduction potential may be obscured.

2.3. Comparison of Digitalization Research Methods

With scholars’ research on enterprise digitalization, many digitalization measurement methods have also emerged, such as the expert scoring method, text analysis method, and digitalization project input amount. One of the mainstream methods is the text analysis method, which is widely used in the measurement of the degree of enterprise digitalization. It extracts the digitalization keywords in the annual reports of enterprises and analyzes the frequency of these words to derive the enterprise digitalization index [17]. However, this method is controversial in terms of accuracy, and it is difficult to take into account the actual investment of enterprises in digitalization technology.

The other research method is the digitalization investment ratio. Enterprises’ digital input is an important indicator of the degree of digitalization, and Liu found in his study that enterprises with a high percentage of digital input are more inclined to adopt intelligent equipment and energy-saving technologies, thus achieving lower carbon emissions [18]. This study plans to analyze the degree of digitalization of enterprises by using the digitalization part of their intangible asset input species.

According to the classification standards of the Bureau of Statistics, the manufacturing industry is divided into a total of eighteen sectors. The level of digitalization varies significantly across industries and regions. Taking Guangdong Province as an example, the application of digital technology differs notably among the textile industry, the refining industry, and the electronics industry. In the textile industry, many enterprises have adopted intelligent system such as “shuttle intelligent textiles”, enabling digital technologies to cover the entire process from business operations to production. However, the textile industry in Guangdong is mainly composed of small and scattered micro-enterprises, which makes it difficult to achieve centralized, large-scale production. This structural feature has limited the widespread adoption of digital technologies in the sector [19]. Moreover, in the high-carbon-emissions sector, there is a focus on automated production in the refining sector, but a wide range of digital twin technology is also being applied to the production chain, reliant on the whole domain data sharing and intelligent model, to build a series of digital scheduling commands, intelligent material balance, and other digital intelligence application scenarios [20]. In addition to these two types of industries, the original information technology is also accounted for as a larger share of the electronics industry, such as the electronic devices and communications equipment industry; this is the basis of industrial digitalization, in which a large number of 5G base stations are being constructed, the Internet of Things is being managed, and so on. The infrastructure in the development of these industries continues to improve, but these industries are also the focus of government policy to support the objective.

In summary, while the overall impact of industrial digitalization on carbon emissions has been studied to some degree, the literature still lacks a detailed differential analysis at both the regional and sector levels. Most studies focus on macro-level emission reduction effects, with insufficient exploration of the differences among specific regions or specific industries.

Based on the existing literature, this study aims to address the following research gaps:

• It delves into the correlation between digitalization investments and carbon emissions among listed manufacturing firms in Guangdong Province, thereby furnishing empirical evidence to underpin regional-level carbon emissions research.
• It explores the heterogeneity of the carbon reduction effects across different industries during the digital transformation process, providing valuable insights to refine policy development and enhance resource allocation strategies.

3. Model Building

3.1. Digitalization Degree Calculation

To measure the degree of digital transformation, this study chooses an improved version of the word frequency method by Li et al. [18]. It compares the investment amounts disclosed in the special reports on annual fundraising and the construction in progress and intangible assets (digital part) in the financial statements. If keywords or relevant patents related to digital transformation technology, such as “software” and “network”, are mentioned in the breakdown of the financial report, the item will be defined as digital technology intangible assets. Cumulative investment in digital transformation was derived and logged. The digitalization data are derived from the publicly released financial statements of the companies.

3.2. Carbon Emission Accounting

The environmentally-extended input–output (EEIO) model is an extended environmental input–output model that introduces environmental elements into the input–output model.

At the first step, this research constructs the technical coefficient matrix $A_{ij}$, where $Z_{ij}$ is the intermediate input of sector i to sector j, and $X_j$ represents the total output of sector j.

$$A_{ij}={\displaystyle \frac{Z_{ij}}{X_j}}$$

(1)

Then, we construct the base total output ${\mathrm{X}}_0$ and base carbon emission ${\mathrm{C}}_0$. In this formula, Y is the final demand, and F is the carbon emission factor, which is derived from the ratio between carbon emissions and sectoral output.

$$X={\left(I-A\right)}^{-1}Y$$

(2)

$$C=FX=F{\left(I-A\right)}^{-1}Y$$

(3)

This paper mainly focuses on the impact of digitalization within the manufacturing sector, while the input–output table contains other industry sectors except manufacturing, in order to reduce the impact; thus, “agriculture, forestry, animal husbandry, and fishery” and other sectors are merged into “other sectors”, and the remaining 17 sectors are studied individually. Since the carbon intensity of each sector varies, for example, the carbon emissions of the metal manufacturing sector are much higher than those of other sectors, the sectors are subdivided into high-carbon and non-high-carbon sectors according to whether the carbon intensity is greater than 0.1. High-carbon sectors include “petroleum, coking products, and processed nuclear fuel”, “nonmetal mineral products”, and “smelting and pressing of metals”.

3.3. Data Sources

In this study, the input–output data are from the 2017 Guangdong Province Input–Output Tables publicly released by China’s statistics department, which are published every five years, with 2017 being the most recent data available. Emissions from various sectors in Guangdong Province are mainly from the China Carbon Accounting Databases (CEADs), which provide statistics on carbon emissions from various sectors in Guangdong Province in 2017.

3.4. SDA Model

The structural decomposition analysis (SDA) method is often used in conjunction with input–output modeling, which is based on input–output tables and is widely used to examine the impact of drivers on carbon emissions [21]. Decomposing carbon emissions into technological, scale, and structural effects forms the core analytical framework of the SDA model. As this study examines the impact of digital technology adoption—particularly in production, transportation, and related processes—on carbon emissions in the manufacturing sector, a simplified SDA approach is employed to estimate the contribution of technological advancement, including digitalization, to emission reductions. This method allows for a clearer identification of the role of technological factors in shaping emission trends and provides empirical evidence to support the environmental benefits of digital transformation in manufacturing.

The SDA model usually takes two forms: weighted average method and bipolar decomposition method. In order to simplify the arithmetic steps, reduce the error, and obtain more reasonable results, this paper adopts the bipolar decomposition method to decompose the change ΔC of carbon emissions in the two periods of 2012 and 2017 into the technological effect, structural effect, and scale effect and to explore the role of the influence of different factors.

The basic form of the SDA model is:

$$S=BY$$

(4)

$$B={\left(I-A\right)}^{-1}$$

(5)

From the EEIO model, we have:

$$C=FX=F{\left(I-A\right)}^{-1}Y$$

(6)

Setting 2012 as the base period (period 0) and 2017 as the comparison period (period 1):

$$C_0=F_0{B}_0{Y}_0$$

(7)

$$C_1=F_1{B}_1{Y}_1$$

(8)

Therefore, the change in total carbon emissions is assumed to be:

$$\Delta C=C_0-C_1$$

(9)

Based on the polarized average method, the decomposed form of the impact of digitalization on carbon emissions in different sectors of the manufacturing industry can be written as follows:

$$\Delta C={\displaystyle \frac{1}{2}}{\Delta F(B}_0{Y}_0+B_1{Y}_1)+{\displaystyle \frac{1}{2}}\left(F_1\Delta B{Y}_0+F_0\Delta B{Y}_0\right)+{\displaystyle \frac{1}{2}}\left(F_0{B}_0+F_1{B}_1\right)\Delta Y$$

(10)

Within the formula, $\Delta \mathrm{F}$ reflects technological effects, that is, the changes in carbon intensity per unit of production value, $\Delta \mathrm{B}$ reflects the structural change, and $\Delta \mathrm{Y}$ reflects the scale change.

The tech effect, struct effect, and scale effect can be written as follows:

$$TechEffect={\displaystyle \frac{1}{2}}{\Delta F(B}_0{Y}_0+B_1{Y}_1)$$

(11)

$$StructEffect={\displaystyle \frac{1}{2}}\left(F_1\Delta B{Y}_0+F_0\Delta B{Y}_0\right)$$

(12)

$$ScaleEffect={\displaystyle \frac{1}{2}}\left(F_0{B}_0+F_1{B}_1\right)\Delta Y$$

(13)

The technology effect mainly measures the extent to which digitalization or other technological improvements reduce or increase carbon emissions per unit of output value, while the structure effect reflects the impact of changes in the input–output relationship between industries (such as changes in the Leontief inverse matrix B) on carbon emissions. This effect mainly reflects the changes in carbon emissions due to industrial restructuring, including direct and indirect (pull and push) effects, and the scale effect reflects the impact of changes in final demand or economic scale on carbon emissions, i.e., the increase in carbon emissions brought about by the expansion of economic activities.

As we can see in

Table 1. Result of SDA model.

	Tech Effect (Tons)	Struct Effect (Tons)	Scale Effect (Tons)
Food and Tobacco Processing	−1,523,954.663	−40,273.39793	1,554,247.136
Textile Industry	−6,334,503.097	2,743,807.047	4,009,767.908
Textile Clothing Shoes and Hats Leather Down and Its Products	−1,265,568.841	−211,542.9817	940,001.822
Woodwork and Furniture	−474,611.8052	189,779.7323	290,159.3434
Paper, Printing, and Stationery	−3,456,984.744	2,259,874.772	2,784,573.23
Petroleum, Coking Products, and Processed Nuclear Fuel	4,350,120.727	−3,572,788.784	1,543,157.142
Chemical Products	−414,517.3999	733,783.2996	2,764,731.79
Nonmetal Mineral Products	−38,760,022.87	87,124,698.11	70,779,646.35
Smelting and Pressing of Metals	7,675,511.848	76,891,809.72	126,129,635.2
Metal Products	−1,622,959.736	335,345.0404	1,148,140.695
Ordinary Machinery	−120,892.5008	140,086.1948	1,218,281.694
Equipment for Special Purposes	−622,493.426	−232,772.5191	593,749.7327
Transportation Equipment	−768,201.821	−166,827.9607	595,702.0655
Electric Equipment and Machinery	−1,302,721.86	−234,572.8099	809,394.5448
Telecommunications Equipment, Computers, and Other Electronic Equipment	−3,414,791.295	−1,749,447.061	2,310,817.422
Instrumentation Machinery	−415,960.8536	129,899.6814	210,468.0162
Other Manufacturing Products and Scrap Waste	−58,676,734.6	47,202,190.78	8,200,623.011

, the results reveal that industries such as food and tobacco processing within the manufacturing sector exhibit negative values under the technology effect category, indicating a decline in carbon intensity per unit of output during the study period as a result of technological progress and digital transformation. For instance, the food and tobacco processing industry achieved a reduction of 1.52 million tons of carbon emissions through improvements in energy efficiency, facilitated by automation and energy management systems. Similarly, the textile and furniture manufacturing industries recorded reductions of 6.33 million tons and 0.47 million tons, respectively. Under Guangdong Province’s Industrial Big Data Integration and Application Project, the chemical engineering sector cut carbon emissions by 0.41 million tons. By deploying a third-party data platform to optimize energy management and process workflows, the project notably enhanced the technology-driven reduction of emissions. These findings show that even traditionally labor-intensive sectors hold considerable potential for energy conservation and emissions reduction through digitalization.

However, the structural and scale effects exhibited a different situation. Certain industries show positive structural effects, indicating that the contribution of industrial restructuring to emission reductions remains limited. Meanwhile, scale effects were predominantly positive across sectors, implying that the adoption of digital technologies may have driven industrial expansion and higher output, thereby contributing to an increase in carbon emissions. Additionally, the expansion of production demand has offset part of the gains achieved through technological improvements. This underscores that under conditions of continued economic growth, relying solely on end-of-pipe pollution control is insufficient to meet carbon neutrality targets. Instead, greater emphasis must be placed on reducing emissions at the source through front-end digital upgrades and optimization of the energy mix.

3.5. Scenario Simulation

In this paper, I use a scenario simulation approach to examine how a static input–output model affects the digitalization process in an industry subject to external shocks. In their investigation of the economic and carbon emission impacts associated with electric vehicle rollout, Guo et al. [22] employed a similar simulation strategy to replicate carbon emission shocks arising from fluctuations in electric vehicle demand. Given the sustained growth of digitalization and its influence on industrial energy consumption, the model is applied to analyze the impact of industrial digitalization on carbon emissions across various manufacturing sectors.

In the base scenario, the digitalization degree is kept constant, and the base carbon emission is ${\mathrm{C}}_0$.

$$C_0=F{X}_0=F{\left(I-A\right)}^{-1}Y$$

(14)

Next, the degree of digital impact, denoted as R, is calculated. Let ${\mathrm{D}}_0$ represent the digitalization rate in 2012 and ${\mathrm{D}}_1$ the rate in 2017. The change in digitalization across industries $\Delta \mathrm{D}$ is first calculated using the following formula. Finally, $\mathrm{R}$ is defined as the average degree of $\Delta \mathrm{D}$.

$$\Delta D={\displaystyle \frac{D_1-D_0}{D_0}}$$

(15)

In Scenario A, digitalization increased by 0.2%. In this scenario, the adjusted gross output ${\mathrm{X}}_1$ and carbon emissions ${\mathrm{C}}_1$ are expressed as the following model:

$$X_1={\left(I-A^{\prime }\right)}^{-1}Y$$

(16)

$$C_1=F_1{X}_1$$

(17)

Adjustment of the technical coefficient matrix to simulate digitally optimized production structures and reduce carbon emissions according to the ${\mathrm{R}}_{\mathrm{i}\mathrm{j}}$ (adjustment rate of 0.2%) is performed:

$$A_{ij}^{\prime }=A_{ij}\left(1-R\right)$$

(18)

The total output and carbon emissions are recalculated according to the model. Then, the carbon emissions gap is analyzed.

The emission reductions of each sector are given as follows:

$$\Delta C_i=C_{0,i}-C_{1,i}$$

(19)

The sectoral contribution rates, which are used to measure the relative contribution of each sector to the total emission reductions, are calculated as follows:

$${Contributionrate}_i={\displaystyle \frac{\Delta C_i}{\Delta C}}$$

(20)

3.6. Result

Guangdong manufacturing industry. Among the high-carbon industries, as shown in the Figure 3 and Table S1, the metal smelting and rolling industry has a contribution of digitalization to the reduction of carbon emissions of 39.69%, which occupies a dominant role in the overall carbon emissions reduction. The nonmetal mineral production industry has a digital reduction contribution of 14.71%. This industry involves the production of building materials such as cement, and these processes are often accompanied by a large amount of carbon emissions, so it has a high carbon emission intensity of up to 0.98. However, the contribution rate is relatively low compared with that of metal smelting, which proves that its digitalization degree needs to be improved and that there is still room for development. Especially in the cement industry, as one of the world’s largest sources of carbon emissions, it is still difficult to reduce emissions, but through the continuous introduction of digital technology and the use of green building materials, it will further reduce the level of emissions.

In contrast, the emission reduction effects in the petroleum, coking, and nuclear fuel industries are relatively limited, with a contribution rate of only around 1.19%. This may be attributed to the low responsiveness of their production processes to digital interventions.

And in the Figure 4 and Table S2, the textile industry and general equipment manufacturing sector have demonstrated moderate carbon reduction effects, contributing approximately 0.37% and 0.56% of the total emission reductions, respectively. These industries are characterized by multiple energy distribution nodes within their production chains, making them particularly suitable for real-time energy monitoring and management. The integration of digital systems in these sectors enables enhanced energy visualization, facilitating more precise and efficient energy control. It is worth noting that some traditional manufacturing sectors, such as the instrumentation machinery industry and the furniture industry, have contributed relatively little to overall carbon emission reductions, accounting for only 0.1% to 0.2%. This limited impact can be attributed, on the one hand, to the inherently low carbon intensity of these sectors and, on the other hand, to their currently low levels of digitalization. These observations suggest that there remains substantial potential for improvement through enhanced adoption of digital technologies in the future.

From the comparative analysis of high-carbon and non-high-carbon industries, although high-carbon industries face greater emission reduction challenges, their carbon emission reductions and contribution rates tend to be larger due to their higher carbon emission factors and the effectiveness of emission reduction measures. Non-high-carbon industries, on the other hand, have relatively low abatement potential, despite their low carbon emission intensity and relatively small contribution rate.

4. Conclusion and Policies

4.1. Research Conclusions and Policy Recommendations

Digitalization represents a major direction for the continued development of the manufacturing industry, and it is essential to achieve energy saving, emission reduction, and production efficiency improvement. However, to ensure a more efficient digital transformation, it is necessary to analyze industry-specific challenges and formulate corresponding development goals and support policies tailored to the characteristics of different sectors. In this study, an EEIO model is employed to examine the positive effects of digital transformation on carbon emissions and industry output in Guangdong Province for the years 2012 and 2017. The contribution rates of digitalization to emission reduction and output growth are calculated separately for each year, allowing for a comparative analysis of how digital transformation has influenced different sectors over time.

(1) The impact of digital transformation is particularly significant in high-carbon sectors.

Carbon emissions from Guangdong’s manufacturing industries are mainly concentrated in the metal smelting and rolling processing and nonmetal mineral products industries. In 2017, the contribution of digitalization to the reduction in carbon emissions in these two industries reached 31.91% and 14.05%, substantially higher than in other sectors. This highlights two key points: first, traditional heavy industry is still the main source of carbon emissions in the manufacturing industry of Guangdong Province; and second, the application of digital technologies has greatly improved the carbon emissions of this type of industry. Therefore, more attention should be devoted to advancing the digital transformation of such sectors through further industrial upgrading, production structure optimization, and targeted policy support, including subsidies for digital technology adoption.

(2) Highly digitalized industries, such as the electronic equipment manufacturing industry, show limited emission reductions but substantial output growth.

Industries such as electronics manufacturing already demonstrate high energy efficiency due to their use of advanced production technologies and automation systems. As a result, digital transformation contributes relatively little to further reductions in carbon emissions per unit of output. However, these industries possess significant growth potential. Digital technologies can greatly enhance production efficiency and streamline management processes, leading to a substantial increase in output value. Additionally, industries such as Guangdong’s chemical sector—especially in pharmaceutical production—already align closely with high digitalization standards and exhibit strong performance in the transition to low-carbon operations.

(3) Textile and other industries are less affected by digitalization.

The textile industry faces challenges introducing advanced digital technologies due to its production characteristics. Operations tend to be decentralized and difficult to standardize, and many enterprises are small in scale. These factors limit the use of automated equipment and smart manufacturing systems, resulting in a generally low level of digitalization. Consequently, the potential for digitalization to enhance production efficiency and reduce carbon emissions per unit of output is weaker, leading to smaller overall emission reduction effects in this sector.

In light of the analysis above, this study offers the following policy recommendations.

(1) Research indicates that digital technologies can effectively reduce carbon emissions by enhancing production efficiency and fostering green innovation. Policymakers should prioritize investments in digital infrastructure and support smart manufacturing upgrades in high-emission industries such as power generation, chemicals, and heavy industry. For example, promoting the construction of industrial internet platforms can facilitate this transformation.

(2) Strengthening the integration of digitalization and green low-carbon technologies in high-carbon industries. For industries such as metal smelting and nonmetallic minerals, digital transformation should be deeply integrated with green low-carbon technologies. Governments should prioritize the application of big data, the Internet of Things (IoT), and other digital solutions to improve energy efficiency, optimize operational processes, and reduce emissions and waste. It is also recommended to develop sector-specific digital platforms incorporating technologies such as the IoT and 5G. These platforms can support real-time monitoring and intelligent control at the production line level and enhance cross-enterprise information exchange. Enterprises themselves should take an active role in adopting digital tools, enhancing their internal technological capabilities, and aligning digital strategies with sustainable development goals.

(3) Support the digital upgrading of low-carbon industries. While high-emission sectors deserve priority, low-carbon industries should not be overlooked. Their production capacity should be expanded, and their role as leaders in green transformation should be reinforced. Advanced digital technologies can improve efficiency and value-added products in these sectors. In parallel, the cultivation of digital talent should be encouraged to drive R&D and the dissemination of innovative technologies. Successful case studies in low-carbon digital transformation can generate replicable models for green upgrading across the broader industrial landscape.

(4) Promoting cross-industry spillover effects of the digital industry. Findings from the structural decomposition analysis and input–output modeling suggest that digital industries not only reduce their own emissions but also contribute to indirect emission reductions in downstream traditional industries via supply chain mechanisms. The government should encourage digital service outsourcing, the development of platform-based enterprises, and deeper collaboration between digital firms and traditional manufacturers to unlock the systemic potential of carbon reduction across the industrial ecosystem.

4.2. Research Limitations and Future Outlook

As the latest version of the 2022 input–output table has not yet been released, this study uses the 2012 and 2017 input–output tables, which fail to cover the latest digitalization change trends. Accordingly, our quantitative estimates of digitalization’s impact on manufacturing carbon emissions may not fully capture the structural and technological advances of recent years, nor the emissions reduction policies implemented since 2017. In the future, the key aspects of digitalization can be tracked dynamically with the latest data of 2022.

Due to data constraints at the provincial-industry level and methodological limits, this study adopts the ratio of intangible assets as a proxy for digitalization. However, this measure may not capture digital maturity as assessed through NLP-driven text mining or enterprise surveys and should be refined in future research.

This study mainly conducts static input–output analysis and lacks a comparison of changes in digitalization in dynamic time. In the future, dynamic input–output models can be used, or linear regression models can be used to assist studies in exploring the long-term impact of digitalization on the industry’s carbon emissions and output value.