Transfer Characteristics of Embodied Carbon Emissions in Export Trade—Evidence from China

Abstract

The export trade of China, the factory of the world, promotes economic growth while increasing carbon emissions. This study integrates China’s multi-regional input–output table and the world input–output table to explore the international transfer-in effect and foreign spillover effect of carbon emissions caused by China’s export trade. A structural decomposition analysis model is also used to identify the influencing factors of carbon emissions caused by China’s export trade of intermediate and final products. Results show that: (1) 45.13–58.87% of the transfer-in carbon emissions resulting from China’s export trade are caused by developed countries and 41.13–54.87% by developing countries; (2) the foreign spillover effect caused by China’s export trade is primarily associated with developing countries, accounting for 63.79–69.61%; (3) carbon emissions caused by the export of intermediate products (final products) in China are primarily caused by the scale effect (industrial linkage). China should adjust the structure of its export trade in accordance with the characteristics of embodied carbon emissions in export trade to achieve low-carbon development.

1. Introduction

China’s exports increased significantly after joining the World Trade Organization in 2001. Compared with 2001, China’s exports increased by 492.92% in 2010. With China’s rapid increase in exports, the ratio of its exports to the world has increased substantially. China’s share of world exports increased from 4.3% in 2001 to 10.31% in 2010, exceeding the United States (8.36%), Germany (8.23%), and Japan (5.03%). Exports have contributed significantly to the growth of China’s economy. As a result of its nominal GDP exceeding that of Japan in 2010, China has become the world’s second-largest economy [1]. Nonetheless, China’s rapidly expanding export trade also contributes to significant amounts of domestic environmental pollution, including carbon emissions [2]. For the development of effective environmental policies, the environmental impact of exports must be considered. In this context, the paper examines the characteristics of the transfer of embodied carbon emissions in China’s export trade.

In energy and environment research, climate change has always been an important topic, among which China’s carbon emissions are included [3,4,5,6,7]. The literature on China’s carbon emissions typically examines environmental regulations, industrial structure, and urbanization as factors affecting emissions [8,9,10,11]. According to the literature, factors such as technological progress and energy restructuring are critical to China’s low-carbon development [12,13,14]. However, systematic research on the characteristics of China’s carbon emissions is lacking from a trade perspective. Accordingly, this study systematically examines the characteristics of carbon emissions in the process of China’s export trade and provides empirical evidence for China to achieve carbon neutrality.

We focus on China for several reasons. First, with the largest export volume in the world, China is ideally suited for studying the characteristics of carbon emissions caused by export trade to achieve a low-carbon economy within China. Second, because China is the world’s largest carbon emitter, paying attention to its low-carbon development is important for achieving carbon neutrality [2]. Third, China is the world’s largest developing country. With China’s success in balancing economic growth with the reduction in carbon emissions, it serves as a model for other developing countries. It also contributes to economic sustainability around the world.

Based on China’s multi-regional input–output table (MRIOT) and the world input–output table (WIOT), this study constructs a multi-regional input–output (MRIO) model for China’s provinces in relation to the other countries and analyzes the international transfer-in effect and foreign spillover effect of carbon emissions caused by China’s export trade. The structural decomposition analysis (SDA) method is also applied to determine the factors influencing carbon emissions resulting from China’s export of intermediate and final products. The purpose of this study is to provide empirical evidence for China’s low-carbon development as well as its attainment of carbon neutrality.

The paper is relevant to the literature on international trade. International trade not only contributes to economic growth but also often causes environmental degradation. Scholars have, therefore, conducted extensive research on international trade from different perspectives. Researchers in economics have examined the effects of international trade on innovation [15,16,17,18,19], productivity [20,21,22,23,24,25], and the intensity of emissions [26,27,28,29]. International trade has a significant impact on carbon emissions [30,31]. International trade is responsible for approximately 25% of global carbon emissions [32]. Scholars of environmental science have studied the transfer characteristics of carbon emissions resulting from trade [2,3,33,34]. According to Weber et al. (2008), exports accounted for about one-third of China’s carbon emissions in 2005 [2]. Literature traditionally treats international trade as a whole based on China’s MRIOT, or China as a whole based on the WIOT. However, the transfer characteristics of carbon emissions caused by China’s provincial export trade to other countries have not been fully studied.

Climate change is a major issue in environmental science, and the study of carbon emission transfer is among the key branches of this field [35]. According to the scope of trade, carbon transfer in trade can be divided into two types, namely, international trade and domestic trade. Scholars began to examine the transfer of carbon emissions in international trade, which frequently occurs between developed and developing countries. These studies have demonstrated that developed countries are responsible for a vast number of carbon emissions that are emitted by developing countries [30,36,37,38,39,40]. Etyuhtul et al. (2016), for example, found that trade liberalization has made China, India, Indonesia, and Turkey pollution havens for developed countries [41]. As trade between developing countries increased, some scholars began to study the impact of trade between developing countries on carbon emissions [42]. Some scholars examined the impact of trade between developed countries on carbon emissions [43].

Other scholars also began to view a country as a heterogeneous whole, and then discussed the carbon emissions caused by domestic inter-regional trade [3,33,34,44,45,46]. Existing literature often discusses the transfer of carbon emissions between countries or between regions within a country. However, it neglects the characteristics of international and provincial transfer of carbon emissions caused by China’s exports, which is not conducive to the implementation of regional responsibility for carbon emissions. This study is particularly relevant to several existing studies [3,32]. Zhou et al. (2018) examined how carbon emissions transferred through major regions using China’s MRIOT, and found emissions predominantly transferred from less developed regions to developed regions [3]. According to Zhao et al. (2014), the WIOT was used to measure carbon emissions embodied in China’s international trade, and the findings indicated that emissions embodied in export and import trade increased significantly during 1995–2009 [32]. This study differs from the above literature in three key respects. First, China’s MRIOT and WIOT are incorporated in this study to examine the carbon transfer caused by China’s provincial export trade to other countries. The literature above is often based merely on China’s MRIOT or WIOT. Second, this study examines not only the foreign spillover effect of emissions resulting from China’s exports but also the international transfer-in effect. Third, the study investigates the factors that affect emissions resulting from the exports of intermediate and final products by using the SDA method.

This research is related to the literature on estimating the influencing factors of carbon emissions. Input–output analysis is widely used in calculating embodied carbon emissions, including two types of models: single-region input–output (SRIO) model [47,48,49] and MRIO [3,36,50,51]. Compared with the SRIO model, the MRIO model takes into account the technical differences across regions and allows tracking of carbon emissions through interregional trade links, which is an effective method of analyzing carbon transfer [52,53]. For instance, Duarte et al. (2018) applied an MRIO model to examine the trajectories of global carbon emissions during 1995–2009 [54]. Existing studies often combine the MRIO model with the SDA method to explore the determinants of carbon emissions [55,56,57]. Deng and Xu (2017), for example, examined the impact of emission intensity, production structure, and final demand on carbon emissions using the SDA method [58]. As opposed to the existing literature, this paper examines the transfer characteristics of carbon emissions resulting from China’s export trade. It also discusses the influencing factors of emissions in the export of intermediate and final products using the SDA method.

Compared with previous studies, this study provides the following contributions. First, this study integrates China’s MRIOT and the WIOT to construct an MRIO model between the provinces of China and other countries during 2007–2012 and to analyze the characteristics of carbon emissions resulting from China’s export trade. Second, this study systematically investigates the carbon transfer in the process of China’s export trade. Specifically, it examines the international transfer-in effect and the spillover effect of emissions caused by China’s exports. Third, this study uses the SDA method to explore the influencing factors of emissions caused by the export trade of intermediate and final products, which is of great importance to the adjustment of export structure and the realization of carbon neutrality.

The rest of this study is as follows. The second part reviews the literature. The third part describes the model and data. The fourth part presents our results. The fifth part provides conclusions and policy implications.

2. Model and Data

By incorporating China’s MRIOT and WIOT (Supplementary Materials), this study constructs a MRIO model for China’s provinces in relation to the other countries and identifies the international transfer-in effect and foreign spillover effect of carbon emissions caused by China’s export trade. The SDA method is also applied for determining the factors that influence China’s carbon emissions from its export of intermediate and final products.

2.1. MRIO Model

Based on the LMRIO table, this study examines the characteristics of carbon transfer in China’s export trade, including the international transfer-in effect and foreign spillover effect caused by export trade. Based on the MRIO model, we can deduce that China’s export trade results in its provincial production.

$$\left(\begin{array}{c}E{m}^{P_1}\\ ⋮\\ E{m}^{P_{30}}\\ E{m}^{D_1}\\ ⋮\\ E{m}^{D_{43}}\end{array}\right)=\left(\begin{array}{cccccc}{L}^{P_1{P}_1}& \cdots & L^{P_1{P}_{30}}& L^{P_1{D}_1}& \cdots & L^{P_1{D}_{43}}\\ ⋮& \ddots & ⋮& ⋮& \ddots & ⋮\\ L^{P_{30}{P}_1}& \cdots & L^{P_{30}{P}_{30}}& L^{P_{30}{D}_1}& \cdots & L^{P_{30}{D}_{43}}\\ L^{D_1{P}_1}& \cdots & L^{D_1{P}_{30}}& L^{D_1{D}_1}& \cdots & L^{D_1{D}_{43}}\\ ⋮& \ddots & ⋮& ⋮& \ddots & ⋮\\ L^{D_{43}{P}_1}& \cdots & L^{D_{43}{P}_{30}}& L^{D_{43}{D}_1}& \cdots & L^{D_{43}{D}_{43}}\end{array}\right)\left(\begin{array}{ccc}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & 0\\ ⋮& \ddots & ⋮\\ 0& \cdots & {{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ 0& \cdots & 0\\ ⋮& \ddots & ⋮\\ 0& \cdots & 0\end{array}\right)$$

(1)

$$=\left(\begin{array}{ccc}{L}^{P_1{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & L^{P_1{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ ⋮& \ddots & ⋮\\ L^{P_{30}{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & L^{P_{30}{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ L^{D_1{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & L^{D_1{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ ⋮& \ddots & ⋮\\ L^{D_{43}{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & L^{D_{43}{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\end{array}\right)$$

$E{m}^p$ on the left side of Equation (1) represents the output matrix resulting from the export trade of P province in China. The right side of Equation (1) is the product of the Leontief inverse matrix and the diagonal matrix of China’s exports vector. According to these production relations, the MRIO model of environmental extension can be stated as follows:

$$T{E}^{PC}=\left(\begin{array}{ccc}{\displaystyle \sum }_{i=43}T{E}^{P_1{D}_i}& \cdots & {\displaystyle \sum }_{i=43}T{E}^{P_{30}{D}_i}\end{array}\right)$$

(2)

$$=\left(\begin{array}{ccc}{f}^{P_1}{L}^{P_1{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & f^{P_1}{L}^{P_1{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ ⋮& \ddots & ⋮\\ f^{P_{30}}{L}^{P_{30}{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & f^{P_{30}}{L}^{P_{30}{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ f^{D_1}{L}^{D_1{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & f^{D_1}{L}^{D_1{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\\ ⋮& \ddots & ⋮\\ f^{D_{30}}{L}^{D_{43}{P}_1}{{\displaystyle \sum }}_{i=43}E{x}^{P_1{D}_i}& \cdots & f^{D_{30}}{L}^{D_{43}{P}_{30}}{{\displaystyle \sum }}_{i=43}E{x}^{P_{30}{D}_i}\end{array}\right)$$

$T{E}^{PC}$ indicates the transfer of emissions resulting from China’s export trade. $P_i$ denotes the provinces of China, i = 1... 30. $D_j$ indicates the rest of the world, j = 1... 43. Taking the export trade of the province of Ps as an example, we decomposed the carbon emissions resulting from its export trade in China.

$$T{E}^{P_s{D}_i}=\underset{\mathrm{International} \mathrm{transfer}-\mathrm{in} \mathrm{effect}}{\underbrace{{{\displaystyle \sum }}_{i=43}{f}^{P_1}{L}^{P_1{P}_s}E{x}^{P_s{D}_i}+{{\displaystyle \sum }}_{i=43}{f}^{P_2}{L}^{P_2{P}_s}E{x}^{P_s{D}_i}+\cdots +{{\displaystyle \sum }}_{i=43}{f}^{P_{30}}{L}^{P_{30}{P}_s}E{x}^{P_s{D}_i}}}$$

(3)

$$+\underset{\mathrm{Foreign} \mathrm{spillover} \mathrm{effect}}{\underbrace{{{\displaystyle \sum }}_{i=43}{f}^{D_1}{L}^{D_1{P}_s}E{x}^{P_s{D}_i}+\cdots +{{\displaystyle \sum }}_{i=43}{f}^{D_{43}}{L}^{D_{43}{P}_s}E{x}^{P_s{D}_i}}}$$

$$=\underset{\mathrm{Local} \mathrm{emission} \mathrm{effect}}{\underbrace{\underset{\left(1\right) \mathrm{Export} \mathrm{of} \mathrm{final} \mathrm{product}}{\underbrace{{{\displaystyle \sum }}_{i=43}{f}^{P_s}{L}^{P_s{P}_s}{Y}^{P_s{D}_i}}}+\underset{\left(2\right) \mathrm{Export} \mathrm{of} \mathrm{intermediate} \mathrm{product}}{\underbrace{{{\displaystyle \sum }}_{i=43}{f}^{P_s}{L}^{P_s{P}_s}{Z}^{P_s{D}_i}}}}}$$

$$+\underset{\mathrm{Domestic} \mathrm{spillover} \mathrm{effect}}{\underbrace{\underset{\left(3\right) \mathrm{Export} \mathrm{of} \mathrm{intermediate} \mathrm{product}}{\underbrace{{{\displaystyle \sum }}_{j\ne s}{{\displaystyle \sum }}_{i=43}{f}^{P_j}{L}^{P_j{P}_s}{Z}^{P_s{D}_i}}}+\cdots +\underset{\left(4\right) \mathrm{Export} \mathrm{of} \mathrm{final} \mathrm{product}}{\underbrace{{{\displaystyle \sum }}_{j\ne s}{{\displaystyle \sum }}_{i=43}{f}^{P_j}{L}^{P_j{P}_s}{Y}^{P_s{D}_i}}}}}$$

$$+\underset{\left(5\right) \mathrm{Foreign} \mathrm{spillover} \mathrm{effect}}{\underbrace{{{\displaystyle \sum }}_{i=43}{f}^{D_1}{L}^{D_1{P}_s}E{x}^{P_s{D}_i}+\cdots +{{\displaystyle \sum }}_{i=43}{f}^{D_{43}}{L}^{D_{43}{P}_s}E{x}^{P_s{D}_i}}}$$

In Equation (3), item (1) indicates that the direct export of final products from China’s $P_s$ province results in carbon emissions in that region. As indicated by item (2), the direct export of intermediate products from China’s province results in carbon emissions in the region. The first two items are collectively known as the “local emission effect.” Item (3) indicates that other provinces in China export their intermediate products through $P_s$ province, resulting in carbon emissions. In item (4), other provinces in China export their final products through $P_s$ province, resulting in carbon emissions. We refer to items (3) and (4) collectively as the “domestic spillover effect.” As stated in item (5), other countries (regions) export their intermediate products through $P_s$ province, resulting in carbon emissions, known as the “foreign spillover effect.”

2.2. SDA Method

The export trade of China is further decomposed into scale effect $\rho$, regional structure effect $\eta$, and product structure effect $\delta$. Consequently, we can derive:

$$E^d=\widehat{F}{\left(I-A\right)}^{-1}Ex=\widehat{F}L\delta \eta \rho$$

(4)

where $E^d$ represents carbon emissions resulting from China’s export trade. The Leontief inverse matrix is denoted as $L={\left(I-A\right)}^{-1}$. $\delta$ represents the product structure of China’s export trade. $\eta$ indicates the regional structure of export trade. China’s total exports are denoted by $\rho$.

For the comparison period, we set the carbon emissions associated with export trade as $E_t^d$, and for the base period, we set them as $E_0^d$. Therefore, the change in emissions caused by exports during the period is $E_0^d$. The symbol “Δ” denotes the change in carbon emissions.

We can decompose the change in $E^d$ using Equation (4) to derive $\mathsf{\Delta }{E}^d$:

$$\mathsf{\Delta }{E}^d=E^d\left(\mathsf{\Delta }\widehat{F}\right)+E^d\left(\mathsf{\Delta }L\right)+E^d\left(\mathsf{\Delta }\delta \right)+E^d\left(\mathsf{\Delta }\eta \right)+E^d\left(\mathsf{\Delta }\rho \right)$$

(5)

As a result of Equation (5), we can identify the impact of five factors on emissions caused by export trade. Given that the SDA form is not unique, the “average of the two polar decompositions” was adopted in this study [59]. The change in carbon emissions caused by exporting intermediate products can be estimated as follows:

$$\mathsf{\Delta }{E}_z^d=\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{t}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{n}\mathrm{o}\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left(\mathsf{\Delta }\widehat{F}{L}^0{\delta }_z^0{\eta }_z^0{\rho }_z^0+\mathsf{\Delta }\widehat{F}{L}^t{\delta }_z^t{\eta }_z^t{\rho }_z^t\right)}}+\underset{\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{l} \mathrm{l}\mathrm{i}\mathrm{n}\mathrm{k}\mathrm{a}\mathrm{g}\mathrm{e}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t\mathsf{\Delta }L{\delta }_z^0{\eta }_z^0{\rho }_z^0+{\widehat{F}}^0\mathsf{\Delta }L{\delta }_z^t{\eta }_z^t{\rho }_z^t\right)}}$$

(6)

$$+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t} \mathrm{s}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t\mathsf{\Delta }{\delta }_z{\eta }_z^0{\rho }_z^0+{\widehat{F}}^0{L}^0\mathsf{\Delta }{\delta }_z{\eta }_z^t{\rho }_z^t\right)}}+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{r}\mathrm{e}\mathrm{g}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{a}\mathrm{l} \mathrm{s}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t{\delta }_z^t\mathsf{\Delta }{\eta }_z{\rho }_z^0+{\widehat{F}}^0{L}^0{\delta }_z^0\mathsf{\Delta }{\eta }_z{\rho }_z^t\right)}}$$

$$+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{s}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t{\delta }_z^t{\eta }_z^t\mathsf{\Delta }{\rho }_z+{\widehat{F}}^0{L}^0{\delta }_z^0{\eta }_z^0\mathsf{\Delta }{\rho }_z\right)}}$$

The change in emissions caused by exporting final products can be calculated as follows:

$$\mathsf{\Delta }{E}_y^d=\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{t}\mathrm{e}\mathrm{c}\mathrm{h}\mathrm{n}\mathrm{o}\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left(\mathsf{\Delta }\widehat{F}{L}^0{\delta }_y^0{\eta }_y^0{\rho }_y^0+\mathsf{\Delta }\widehat{F}{L}^t{\delta }_y^t{\eta }_y^t{\rho }_y^t\right)}}+\underset{\mathrm{I}\mathrm{n}\mathrm{d}\mathrm{u}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{i}\mathrm{a}\mathrm{l} \mathrm{l}\mathrm{i}\mathrm{n}\mathrm{k}\mathrm{a}\mathrm{g}\mathrm{e}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t\mathsf{\Delta }L{\delta }_y^0{\eta }_y^0{\rho }_y^0+{\widehat{F}}^0\mathsf{\Delta }L{\delta }_y^t{\eta }_y^t{\rho }_y^t\right)}}$$

(7)

$$+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t} \mathrm{s}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t\mathsf{\Delta }{\delta }_y{\eta }_y^0{\rho }_y^0+{\widehat{F}}^0{L}^0\mathsf{\Delta }{\delta }_y{\eta }_y^t{\rho }_y^t\right)}}+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{r}\mathrm{e}\mathrm{g}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{a}\mathrm{l} \mathrm{s}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t{\delta }_y^t\mathsf{\Delta }{\eta }_y{\rho }_y^0+{\widehat{F}}^0{L}^0{\delta }_y^0\mathsf{\Delta }{\eta }_y{\rho }_y^t\right)}}$$

$$+\underset{\mathrm{T}\mathrm{h}\mathrm{e} \mathrm{s}\mathrm{c}\mathrm{a}\mathrm{l}\mathrm{e} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t}}{\underbrace{\frac{1}{2}\left({\widehat{F}}^t{L}^t{\delta }_y^t{\eta }_y^t\mathsf{\Delta }{\rho }_y+{\widehat{F}}^0{L}^0{\delta }_y^0{\eta }_y^0\mathsf{\Delta }{\rho }_y\right)}}$$

2.3. Data

Three types of data were involved in the analysis of this study. The first was China’s MRIOT in 2007, 2010, and 2012, which were obtained from the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. From the input–output tables of 2007 and 2010, 30 sectors were consolidated into 17 sectors, while 42 sectors were consolidated into 17 sectors in the input–output tables of 2012. As of 2012, Tibet was incorporated into Xinjiang, ensuring a unified input–output table with the tables of 2007 and 2010. The second type of data were the World Input–Output Database funded by the European Union, whose main component is the WIOT. Currently, two versions have been released. The data released in 2013 covered the economic and trade relations among 35 sectors in 41 countries (regions) from 1995 to 2011, while the data released in 2016 covered the economic and trade relations among 56 sectors in 44 countries (regions) from 2000 to 2014. To match China’s MRIOT, this study adopted the data released in 2016 and merged its 56 sectors into 17 sectors. To conduct the SDA, MRIOT and WIOT for 2010 and 2012 were deflated to constant prices for 2007. The third type of data were carbon emission data, derived from the environment account of the WIOD database released in 2016, which provided China’s annual carbon emissions at the sector level from 2000 to 2016.

3. Results

With the integration of China’s MRIOT and WIOT, this study examines the international transfer-in effect and the foreign spillover effect of carbon emissions caused by China’s export trade of intermediates and final products, and analyzes the influencing factors of emissions.

3.1. Overall Characteristics of Carbon Transfer Caused by China’s Exports

The issue of climate change is global, and export trade is an important component of global economic activity. Therefore, focusing on carbon emissions resulting from exports is of great importance. In

Table 1. Summary of the carbon emissions resulting from export trade, 2007–2012.

	2007	2010	2012
Production-side emissions (PSE)	7099.83	8499.98	9511.12
Total emissions resulting from export trade (TEET)	2186.56	2011.50	2144.92
Ratio a	30.80	23.66	22.55
(1) Transfer-in emissions caused by export trade (ITIE)	1909.26	1754.84	1856.76
Ratio b	87.32	87.24	86.57
Ratio c	26.89	20.65	19.52
Local emission effect	1271.74	1184.18	1332.43
Domestic spillover effect	637.53	570.66	524.33
Ratio d	66.61	67.48	71.76
Ratio e	33.39	32.52	28.24
Transfer-in emissions from the export of intermediate product	1088.94	958.10	1053.68
Transfer-in emissions from the export of final product	820.32	796.75	803.08
Ratio f	57.03	54.60	56.75
Ratio g	42.97	45.40	43.25
(2) Foreign spillover effect (FSE)	277.29	256.66	288.16
Ratio h	12.68	12.76	13.43
Ratio i	3.91	3.02	3.03

Note: Ratio a refers to the ratio of TEET to PSE; Ratio b (c) indicates the ratio of ITIE to TEET(PSE); Ratio d (e) refers to the ratios of local emission effect (domestic spillover effect) to ITIE; Ratio f (g) represents the ratios of transfer-in emissions from the export of intermediate product (transfer-in emissions from the export of final product) to ITIE; Ratio h (i) indicates the ratios of FSE to TEET(PSE); Unit: Mt, %.

, we present data on China’s production-side emissions (PSE), total emissions resulting from export trade (TEET), international transfer-in effect (ITIE), and foreign spillover effect (FSE) in 2007, 2010, and 2012. The purpose of this analysis is to clarify the global impact of carbon transfer caused by China’s export trade. TEET can be divided into two parts, namely, ITIE and FSE. According to the perspective of the emission source, the ITIE can be divided into domestic spillover and local emission effects. According to the perspective of trade heterogeneity, ITIE can be divided into transfer-in emissions caused by the export of intermediate products and transfer-in emissions caused by the export of final products.

PSE displayed an upward trend over the period studied, increasing from 7099.83 Mt in 2007 to 9511.12 Mt in 2012, with a growth of 33.96%. Additionally, the GDP of China increased by 99.41% in 2012 compared with 2007. Although the global financial crisis broke out in 2008, China’s economic growth was not significantly affected by the crisis due to China’s industrial policies, e.g., the CNY 4 trillion stimulus plan. TEET, however, showed a U-shaped trend, falling from 2186.56 Mt in 2007 to 2011.50 Mt in 2010, and rising to 2144.92 Mt in 2012. ITIE (FSE) also exhibited a U-shaped trend, declining from 1909.26 (277.29) Mt in 2007 to 1754.84 (256.66) Mt in 2010, and increasing to 1856.76 (288.16) Mt in 2012. These changes reflect not only changes in export structure and advancements in production technology but also the significant negative effect of the 2008 global financial crisis on China’s export trade, which led to a decline in export trade volume, thereby reducing ITIE and FSE. As a result, the ratio of ITIE (FSE) in PSE decreased from 26.89 (3.91)% in 2007 to 19.52 (3.03)% in 2012.

The ITIE includes the domestic spillover effect and the local emission effect, of which the local emission effect accounts for approximately 66.61–71.76%. In addition, the domestic spillover effect should also be considered. Therefore, regional governance should be responsible for the reduction in carbon emissions, with coordination among regions being considered. The ITIE also includes both transfer-in emissions resulting from the export of intermediate products and final products. Transfer-in emissions caused by the export of intermediate products mainly account for 54.6–57.03%, which is due to the increase in intermediate goods trade caused by production fragmentation. Therefore, transfer-in emissions from exports of intermediate goods dominate. China’s export trade results in the transfer of 86.67–87.32% of carbon emissions to China, whereas 12.68–13.43% of emissions are transferred to countries (regions) upstream of the industrial chain. Therefore, environmental issues in export trade are primarily the responsibility of exporting countries, and they will also have adverse effects on other countries (regions).

As shown in

Table 2. Provincial-level analysis of emissions caused by China’s export trade, 2007–2012.

Region	2007		2010		2012
	ITIE	PSE	ITIE	PSE	ITIE	PSE
Total	1909.26	7099.83	1754.84	8499.98	1856.76	9511.12
Beijing	2.11	2.57	2.13	2.84	5.19	3.44
Tianjin	2.47	1.60	2.18	1.95	2.16	1.85
Hebei	4.98	6.39	4.70	5.44	3.86	5.18
Shanxi	2.01	2.75	1.80	2.61	1.79	2.35
Inner Mongolia	1.56	2.58	1.92	3.14	1.82	2.85
Liaoning	4.10	4.57	3.20	4.19	3.51	4.45
Jilin	0.64	1.19	0.73	1.40	0.54	1.68
Heilongjiang	1.20	2.15	1.24	1.84	1.01	1.69
Shanghai	6.39	3.54	5.19	3.43	5.77	2.84
Jiangsu	13.67	10.19	13.29	9.02	13.49	9.26
Zhejiang	10.10	7.51	9.63	6.49	10.64	6.50
Anhui	1.41	2.20	1.82	2.78	2.14	3.22
Fujian	4.51	3.28	3.99	2.91	4.24	3.12
Jiangxi	0.94	1.80	2.48	2.71	2.21	2.93
Shandong	9.93	9.68	10.92	11.08	8.58	10.25
Henan	4.04	6.92	4.10	6.51	3.44	6.82
Hubei	1.26	2.64	2.00	3.13	1.38	3.12
Hunan	1.46	2.44	1.82	2.88	1.18	2.91
Guangdong	18.28	12.36	16.46	10.32	16.39	9.37
Guangxi	1.41	2.01	1.65	2.19	0.96	1.92
Hainan	0.21	0.30	0.21	0.33	0.24	0.36
Chongqing	0.52	1.19	0.74	1.39	1.24	1.36
Sichuan	1.18	3.01	2.15	3.63	2.45	4.14
Guizhou	1.01	1.45	0.84	1.23	0.94	1.39
Yunnan	1.49	1.54	1.47	1.61	0.98	1.59
Shaanxi	1.02	1.41	1.52	2.12	1.30	2.14
Gansu	0.72	0.98	0.64	1.08	0.77	1.15
Qinghai	0.21	0.37	0.28	0.39	0.18	0.41
Ningxia	0.32	0.48	0.29	0.50	0.48	0.64
Xinjiang	0.86	0.87	0.60	0.86	1.15	1.06

, coastal provinces deeply involved in global value chains, such as Shanghai, Jiangsu, Zhejiang, Guangdong, and Shandong, contribute a relatively large share to China’s carbon transfer-in emissions caused by exports, which remain essentially around 55%. These provinces were mainly affected by the global financial crisis in 2008, and their share of transfer-in emissions decreased from 58.36% in 2007 to 54.87% in 2012. The share of PSE in coastal provinces decreased from 43.29% in 2007 to 38.21% in 2012. Hence, coastal provinces are more affected by the global value chain, explaining why the ratio of ITIE is higher than that of PSE. The ratio of ITIE was about 1.5% in central and inland provinces, such as Hubei and Anhui, but the trend was different. In Anhui, the ratio of ITIE increased from 1.41% in 2007 to 2.14% in 2012, whereas in Hubei, the trend of change showed an inverted U-shaped pattern, with the ratio rising from 1.26% in 2007 to 2% in 2010, but decreasing to 1.38% in 2012. However, the share of PSE in Anhui province has been increasing, from 2.2% in 2007 to 3.22% in 2012, and is significantly higher than the share of ITIE. The share of PSE in Hubei remains at about 3%, which is also higher than the share of ITIE. Both ratios of ITIE and PSE are low in western provinces, but the situation is the same as in central provinces, i.e., the ratio of ITIE is lower than the ratio of PSE. The ratio of ITIE for Gansu Province, for example, decreased from 0.72% in 2007 to 0.64% in 2010, and then increased to 0.77% in 2012. The PSE ratio increased from 0.98% in 2007 to 1.15% in 2012. The share of PSE is always higher than that of ITIE.

3.2. International Transfer-in Effect and Foreign Spillover Effect of Emissions Caused by China’s Exports

This study explores the global impact of carbon transfer resulting from China’s export trade from two perspectives: (1) the country-specific characteristics of the international transfer-in effect, and (2) the country-specific characteristics of carbon transfer caused by the use of intermediate products from other countries in China’s export trade. Figure 1 illustrates that 45.13–58.87% of the ITIE comes from developed countries, and 41.13–54.87% from developing countries. Moreover, countries (regions) of developed countries that contribute to transfer-in emissions are comparatively concentrated, while countries (regions) of developing countries are more dispersed. The United States, Japan, and South Korea are the top three countries transferring carbon emissions to China. Regardless of trade flows of intermediate products or final products, the share of carbon transferred from these three developed countries to China remains stable. As shown in Figure 2, developing countries are the main emitters of FSE, accounting for 63.79–69.61%, and the share of FSE of developing countries is increasing year by year. FSE accounted for 12.68–13.43% of ITIE. Accordingly, China bears the bulk of the environmental costs in the export process. China also purchases raw materials and other resources from other developing countries, resulting in the indirect export of carbon emissions from those developing countries.

3.3. Domestic Spillover Effect and Local Emission Effect of Emissions Caused by China’s Exports

Figure 3 illustrates the characteristics of domestic spillover and local emission effects resulting from China’s export trade. The local emission effect resulting from the export of intermediate products (LEEI), the local emission effect caused by the export of final products (LEEF), the domestic spillover effect resulting from the export of intermediate products (DSEI), and the domestic spillover effect caused by the export of final products (DSEF) all showed an evident decrease in the eastern, central, and western regions. In coastal provinces, such as Shanghai, Jiangsu, Zhejiang, Guangdong, and Shandong, the share of carbon emissions caused by the domestic spillover effect and local emission effect is higher. The local emission effect, however, is more pronounced. The share of the local emission effect is twice that of the domestic spillover effect. Specifically, the share of the local emission effect caused by the export of intermediate products and final products remained about 70%, whereas that of the domestic spillover effect of intermediate products and final products remained about 30%. In the central provinces, the domestic spillover effect is equivalent to that in the coastal provinces. The ratio of domestic spillover effect in Henan and Hebei provinces is also much higher than the ratio of local emission effect in their own provinces. In Henan Province, for example, the ratio of local emission effect remains about 2%, whereas the ratio of domestic spillover effect remains about 8%. In western provinces, the ratio of the domestic spillover effect and the local emission effect is lower, whereas the domestic spillover effect is more prominent. Therefore, emissions caused by exports mainly occur in China’s coastal provinces, which are dominated by the local emission effect. By contrast, inland provinces are dominated by the domestic spillover effect.

Figure 4 and Figure 5, respectively, illustrate the characteristics of carbon transfer at the provincial level in China resulting from the exports of intermediate and final products in 2012. As shown in Figure 4 and Figure 5, carbon emissions caused by the exports of intermediate products and final products show the same flow characteristics, that is, emissions are transferred from developed regions to less-developed regions, which is in line with previous literature [3,60]. This is because the products produced by regions in China are exported through coastal provinces. The ratio of the transfer-in emissions caused by the export trade of intermediate products (final products) in Shanghai, Jiangsu, Zhejiang, and Guangdong is 44.41–67.49% (49.3–70.8%). This is not only due to the geographical location but also closely related to the industrial structure of these provinces. The central provinces are the main regions of transfer-in emissions caused by export trade in coastal provinces. Specifically, Henan’s export of intermediate (final) products to coastal provinces accounts for 6.45–10.42% (6.2–10.58%) of the domestic spillover effect, while Hebei’s export of intermediate (final) products to coastal provinces accounts for 5.96–11.37% (6.24–11.42%) of the domestic spillover effect.

To further analyze the industrial characteristics of emissions caused by export trade, Figure 6a and Figure 7a, respectively, describe the transfer-in characteristics of domestic spillover effect in representative inland provinces and the transfer-out characteristics of domestic spillover effect in representative coastal provinces in 2012. Figure 6a indicates that metals, electricity, and non-metallic minerals are the sectors with the highest transfer-in emissions for representative inland provinces in China. Jiangxi and Henan, for example, are rich in non-metallic mineral resources, contributing to high emissions from non-metallic minerals. With the exception of sectors with their advantages in some provinces, the sectors with transfer-in emissions in representative inland provinces are mainly concentrated in energy-intensive industries such as metals and electricity. Figure 7a demonstrates that electrical and optical equipment is the sector with the highest transfer-out emissions in China’s representative coastal provinces. Furthermore, the textile and clothes sector in Guangdong and Zhejiang, as well as metals and general equipment in Jiangsu and Zhejiang, are high emitters. Therefore, inland provinces provide high-emission intermediates to coastal provinces through the domestic value chain.

Figure 6b illustrates the industrial characteristics of the local emission effect transfer-in emissions in representative inland provinces in 2012. Consistent with the industrial characteristics of the domestic spillover effect transfer-in emissions, the sectors of the local emission effect transfer-in emissions are also concentrated in metals and electricity. The difference is that the local emission effect of the chemical industry has a particularly high ratio of transfer-in emissions. Furthermore, Inner Mongolia’s transport, postal, and warehousing sector is also a major source of transfer-in emissions due to its unique resources and geographical advantages. In Figure 7b, we illustrate the characteristics of the industrial emissions in representative coastal provinces of the local emission effect transfer-in emissions in 2012. In representative coastal provinces, the sectors of local emission effect transfer-in emissions are primarily metals and electricity. Notably, sectors of the local emission effect transfer-in emissions are concentrated in resource-intensive sectors, whereas the sectors of the domestic spillover effect transfer-out emissions are primarily found in capital-intensive or technology-intensive sectors.

3.4. Structural Decomposition Analysis

We apply the SDA method for determining the factors that influence China’s carbon emissions from the export of intermediate and final products.

3.4.1. Structural Decomposition Analysis of Carbon Emissions Caused by Exports of Intermediate and Final Products

Figure 8 and Figure 9 illustrate the impact of five factors on emissions resulting from the export of intermediate and final products, respectively. The transfer-in emissions resulting from China’s export trade showed a U-shaped trend, decreasing from 1909.26 Mt in 2007 to 1754.84 Mt in 2010, and rebounding to 1856.76 Mt in 2012. In general, the net change in emissions is small because the effects of different factors on carbon emissions are largely canceled out. Consequently, the ratio of changes in transfer-in emissions caused by different factors to changes in net emissions is much greater than 1. The results indicate that industrial linkage and scale effects are the primary driving forces behind the growth in carbon emissions resulting from China’s export trade. Specifically, the scale effect (industrial linkage) is the main driving factor of the growth of emissions caused by China’s export of intermediate products (final products). Relative to the change in net emissions, the scale effect (industrial linkage) contributed to an increase of 464.68% (406.94%) in emissions caused by China’s export of intermediate products (final products) during 2007–2012.

In line with previous studies, the technological effect is the main constraint on the growth of emissions resulting from export trade, whether it is for intermediate or final product exports [61]. The decline in carbon intensity resulted in a 531.5% (875.35%) reduction in emissions from exports of intermediate (final) products during 2007–2012, in comparison with the change in net emissions. The product structure effect and regional structure effect do not appear to have a significant impact on China’s carbon emissions. The regional structural effect, for example, resulted in a reduction in 8.15 Mt (4.71 Mt) in carbon emissions from exports of intermediate (final) products, accounting for 23.12% (27.32%) of the change in net emissions. Although the Chinese government implemented some economic stimulus measures to sustain economic growth during the global financial crisis in 2008, the export trade was still affected, resulting in the emissions caused by the exports to not significantly increase.

3.4.2. Identifying the Factors That Influence Emissions in the Key Sectors

This study further explores industry-level factors that contribute to the change in emissions caused by export trade. In

Table 3. Structural decomposition of industry-level emissions caused by the export trade of intermediate products, 2007–2012.

2007–2010	Primary Industry	Secondary Industry				Tertiary Industry	Total
		Total	Electricity	Metals	Non-Metallic Minerals
Total change	−0.15	−118.23	−78.26	−36.05	−4.89	−12.46	−130.84
Ratio of total change	0.12	90.36	66.19	30.49	4.14	9.53	100
2010–2012	Primary Industry	Secondary Industry				Tertiary Industry	Total
		Total	Electricity	Chemicals	Other manufacturing
Total change	−1.15	84.87	98.52	3.33	1.78	11.86	95.58
Ratio of total change	−1.20	88.79	116.08	3.93	2.09	12.41	100

and

Table 4. Structural decomposition of industry-level emissions caused by the export trade of final products, 2007–2012.

2007–2010	Primary Industry	Secondary Industry				Tertiary Industry	Total
		Total	Electricity	Paper Printing	General Equipment
Total change	−1.33	−15.40	−35.14	−8.35	−2.77	−6.85	−23.58
Ratio of total change	5.62	65.33	228.14	54.18	18.01	29.05	100
2010–2012	Primary Industry	Secondary Industry				Tertiary Industry	Total
		Total	Electricity	Chemicals	Construction
Total change	−3.25	4.69	56.98	0.37	−0.10	4.90	6.33
Ratio of total change	−51.38	74.03	1215.26	7.93	−2.10	77.35	100

, we present the changes in carbon emissions of three industries, caused by the exports of intermediate products and final products in China from 2007 to 2012, respectively. Carbon emissions from China’s export of intermediate products decreased by 130.84 Mt from 2007 to 2010 and increased by 95.58 Mt from 2010 to 2012. The change in emissions caused by China’s intermediate exports (2007–2010, 90.36%; 2010–2012, 88.79%) was largely attributable to secondary industry, as opposed to tertiary industry (2007–2010, 9.53%; 2010–2012, 12.41%) and primary industry (2007–2010, 0.12%; 2010–2012, 1.2%). Therefore, secondary industry was the main industry driving the change in carbon emissions resulting from China’s export of intermediate products. From 2007 to 2010, the ratio of electricity, metals, and non-metallic minerals accounted for 66.19%, 30.49%, and 4.14% of changes in secondary industry emissions, respectively. From 2010 to 2012, the sectors driving the increase in carbon emissions in the secondary industry were electricity, chemicals, and other manufacturing, which accounted for 116.08%, 3.93%, and 2.09% of change in the emissions of the secondary industry, respectively. Accordingly, electricity was the main sector in the secondary industry that affected the change in emissions caused by the export of intermediate products.

The emissions from China’s export of final products decreased by 23.58 Mt from 2007 to 2010, and increased by 6.33 Mt from 2010 to 2012. Most of the change in carbon emissions from China’s export of final products (2007–2010, 65.33%; 2010–2012, 74.03%) also came from secondary industry. However, tertiary industry (2007–2010, 29.05%; 2010–2012, 77.35%) and primary industry (2007–2010, 5.62%; 2010–2012, 51.38%) had a significant influence on the change in carbon emissions caused by China’s final product export trade. The reason was that the impact of each sector on the change in emissions offset the other, resulting in a small change in net emissions caused by the export of the final products. From 2007 to 2010, the ratio of electricity, paper printing, and general equipment accounted for 228.14%, 54.18%, and 18.01% of the change in the emissions of the secondary industry, respectively. From 2010 to 2012, the sectors driving the increase in China’s carbon emissions in the secondary industry were electricity, chemicals, and construction, which accounted for 1215.26%, 7.93%, and −2.1% of the change in the emissions of the secondary industry, respectively. Furthermore, electricity was the primary sector in the secondary industry that affected the change in carbon emissions resulting from exports of final products. This analysis indicates that the key sectors of change in carbon emission caused by export trade have changed over time, but electricity has a dominant role in the change of emissions, which is related to the industrial structure and energy structure.

After determining the key sectors of emissions change caused by China’s export trade, the study further carries out the SDA analysis on the above key sectors as shown in

Table 5. Structural decomposition of carbon emissions caused from key sectors by the export trade of intermediate products, 2007–2012.

Year	Sector	Total Change	Technological Effect	Industrial Linkage	Product Structural Effect	Regional Structural Effect	Scale Effect
2007–2010	Electricity	−78.26	−38.25	−38.19	−7.23	−2.02	7.44
		59.81	29.24	29.19	5.53	1.55	−5.69
	Metals	−36.05	2.18	4.34	−43.94	−2.90	4.28
		27.55	−1.67	−3.32	33.59	2.22	−3.27
	Non-metallic minerals	−4.89	−39.81	15.04	18.80	−1.97	3.05
		3.74	30.43	−11.49	−14.37	1.51	−2.33
2010–2012	Electricity	98.52	8.24	41.16	−2.59	0.13	51.58
		103.08	8.62	43.06	−2.71	0.14	53.97
	Chemicals	3.33	−16.25	9.80	−13.46	−0.53	23.78
		3.49	−17.00	10.25	−14.09	−0.55	24.88
	Other manufacturing	1.78	2.50	−1.01	−0.70	−0.04	1.02
		1.86	2.62	−1.06	−0.73	−0.04	1.07

and

Table 6. Structural decomposition of carbon emissions caused from key sectors by the export trade of final products, 2007–2012.

Year	Sector	Total Change	Technological Effect	Industrial Linkage	Product Structural Effect	Regional Structural Effect	Scale Effect
2007–2010	Electricity	−35.14	−33.30	−16.58	9.67	−0.04	5.11
		149.04	141.25	70.33	−41.03	0.16	−21.67
	Paper printing	−8.35	−10.99	5.97	−3.83	−0.12	0.62
		35.40	46.60	−25.34	16.26	0.50	−2.62
	General equipment	−2.77	−4.88	1.42	0.40	0.15	0.14
		11.77	20.70	−6.01	−1.70	−0.63	−0.61
2010–2012	Electricity	56.98	7.20	35.35	−2.58	−1.18	18.19
		899.61	113.72	558.08	−40.80	−18.57	287.19
	Chemicals	0.37	−10.59	10.74	−5.36	−0.64	6.21
		5.87	−167.14	169.61	−84.56	−10.15	98.11
	Construction	−0.10	−0.02	−0.06	−0.03	0.00	0.01
		−1.55	−0.39	−0.89	−0.49	0.06	0.17

. In different sectors, different factors affect the change in emissions caused by exports. From the perspective of export trade of intermediate products, technological effect and industrial linkage reduced carbon emissions in electricity by 38.25 and 38.19 Mt from 2007 to 2010, respectively. However, the increase in emissions was primarily due to the scale effect. Therefore, technological effect and industrial linkage at this stage were the main factors for the reduction in carbon emissions from electricity due to the export of intermediate products. In contrast with electricity, the product structure effect was the main factor that impacted the reduction in emissions from metals as a result of exporting intermediate products. For non-metallic minerals, the net reduction in carbon emissions resulting from the export of intermediate products was relatively small at 4.89 Mt, as the reduction in emissions resulting from the technological effect was offset by the increase in emissions due to the industrial linkage and product structure effect. During 2010–2012, the factors preventing the increase in carbon emissions in various sectors changed, with net carbon emissions increasing. The main factor that inhibited the increase in emissions in electricity was the product structure effect, but this effect was only 2.59 Mt. No significant increase occurred in the carbon emissions caused by exporting intermediate products because the reduction in the carbon emissions resulting from the technological effect and the product structure effect was offset by the increase in carbon emissions resulting from the industrial linkage and scale effect.

With regard to exports of final products, since the net carbon emissions resulting from China’s final product export trade were relatively small during 2007–2012, the ratio of emissions resulting from various factors to net carbon emissions was relatively high. Although changes occurred in major sectors of emissions resulting from exports of final products, electricity was the most significant sector due to the relatively small impact of changes in emissions resulting from exports of final products in other sectors. For example, between 2007 and 2010, the export of final products from the paper printing reduced carbon emissions by only 8.35 Mt, while between 2010 and 2012, the export of final products from the chemical industry increased carbon emissions by only 0.37 Mt. By analyzing the factors affecting carbon emissions in key industries caused by export trade, we can expect to guide key sectors to achieve “low emission” development through technological progress and improvement of the intermediate input structure.

4. Conclusions and Policy Implications

In this study, we constructed a MRIO model for China’s provinces in relation to the other countries and identified the international transfer-in effect and foreign spillover effect of carbon emissions caused by China’s export trade. The SDA method was also used to determine the factors that affect China’s carbon emissions associated with its exports of intermediate and final products. The study provides empirical evidence and policy implications in support of China’s low-carbon development and achievement of carbon neutrality.

4.1. Conclusions

Based on the integration of China’s MRIOT and WIOT, this study examined the international transfer-in effect and the foreign spillover effect of carbon emissions caused by China’s export trade and analyzed the influencing factors of emissions caused by China’s export trade of intermediate and final goods. The study provides empirical evidence for the adjustment of China’s export trade structure and sustainable economic development. These research conclusions can be summarized as follows.

(1)

Over the study period, China’s production-side emissions increased from 7099.83 Mt in 2007 to 9511.12 Mt in 2012, representing a 33.96% increase. The international transfer-in effect (foreign spillover effect) caused by export trade showed a U-shaped trend, decreasing from 1909.26 Mt (277.29 Mt) in 2007 to 1754.84 Mt (256.66 Mt) in 2010, and rising to 1856.76 Mt (288.16 Mt) in 2012;

(2)

The international transfer of carbon emissions caused by China’s export trade is 45.13–58.87% from developed countries and 41.13–54.87% from developing countries. Additionally, countries (regions) of developed countries contributing to transfer-in emissions are comparatively concentrated, whereas countries (regions) of developing countries are more dispersed. The foreign spillover effect caused by China’s export trade is primarily associated with developing countries, accounting for 63.79–69.61%. Accordingly, environmental issues in export trade are primarily the responsibility of exporting countries and may also have adverse effects on other countries (regions);

(3)

In coastal provinces, such as Shanghai, Jiangsu, Zhejiang, Guangdong, and Shandong, the local emission effect and domestic spillover effect are of greater significance. However, the local emission effect is more pronounced. The effect of domestic spillover is most pronounced in the inland provinces. Carbon emissions derived from the export of intermediate products and final products exhibit the same flow characteristics. This is because most of the products produced by Chinese regions are exported through coastal provinces;

(4)

The results of SDA show that the scale effect (industrial linkage) is the main factor for the growth of emissions resulting from China’s export of intermediate products (final products). Carbon emissions caused by export trade are primarily constrained by the technological effect. Electricity is the primary sector in secondary industry that affects the change in carbon emissions resulting from exports of intermediate and final products.

In terms of limitations, this study examined the transfer characteristics of emissions resulting from China’s export trade. The characteristics of carbon emissions resulting from exports by other countries may vary widely, especially for developed countries. Given that China is the world’s largest developing country and the world’s largest exporter, this fact should not be overlooked. This study also found that the increase in emissions caused by the export of intermediate products (final products) in China was primarily the result of the scale effect (industrial linkage), which may have some involvement with China’s industrial structure and export structure during that period. In light of the transformation of China’s economic structure and the adjustment of its trade structure, the characteristics of the transfer of carbon emissions caused by its export trade may change. Further analyses will be conducted in order to propose appropriate policy recommendations to policymakers.

4.2. Policy Implications

Given that China has become the factory of the world, its export trade has contributed to many environmental problems while promoting economic growth. Using the MRIO model, this study examined the international transfer-in effect and foreign spillover effect of emissions resulting from China’s exports. We further discussed the drivers of carbon emissions generated by China’s export trade by utilizing the SDA method. We propose the following policy implications based on the results of this study. International cooperation to achieve carbon neutrality needs to be strengthened as soon as possible. We found that the international transfer of emissions caused by China’s export trade was 45.13–58.87% from developed countries and 41.13–54.87% from developing countries. In addition, the foreign spillover effect caused by China’s export trade was primarily associated with developing countries, accounting for 63.79–69.61%. This is evidence that a country’s export trade not only increases carbon emissions within its borders but also increases emissions of other countries throughout the global value chains. As a consequence, in responding to climate change, countries should develop stronger cooperation on emission reduction, to avoid carbon leakage and achieve carbon neutrality as soon as possible [62,63]. In addition, China should adjust the structure of its export trade in accordance with the characteristics of carbon emissions. Low-carbon development can be achieved through technological innovation. In this study, we demonstrated that the scale effect (industrial linkage) was the main factor for the growth of emissions caused by China’s export of intermediate products (final products). However, the technology effect dominates the limitations on emissions caused by exports. The factors that drive carbon emissions will shift with economic development, and the predominant factor that restrains the increase in emissions is always the technological effect. Consequently, technological innovation is essential for reducing carbon dioxide emissions [64,65].