Will Trade Protection Trigger a Surge in Investment-Related CO₂ Emissions? Evidence from Multi-Regional Input–Output Model

Abstract

Current research has revealed that global trade promotes transnational investment and contributes to large amounts of CO₂ emissions. Recently, trade protectionism has gradually emerged. This study aims to explore the effect of global trade protection on investment-related CO₂ emissions. We construct an accounting framework for investment-related CO₂ emissions under trade freedom and trade restriction based on the MRIO model for the first time, and investment-related CO₂ emissions in 16 economies are determined in both trade freedom and trade restriction scenarios. The study uses normal trade and no-trade scenarios instead of free trade and restricted trade scenarios. Then, based on the comparison of the two scenarios, the effect of global trade protection on investment-related CO₂ emissions is revealed from the three levels of country, section, and trade links. It is found that global trade protection would lead to an increase of up to 546.17 million tons in investment-related CO₂ emissions under current trade structures. Transnational investment in the trade of end-stage intermediary goods contributed 81.6% of the total effect. In terms of countries, trade protection is quite disadvantageous to CO₂ emission reduction in China and India, and their CO₂ emissions would respectively increase by 105 million tons and 141.5 million tons compared to normal trade. The electricity, gas, and water supply sectors and the manufacturing sector are the main sectors for investment-related CO₂ emissions. This study reveals the effect of trade freedom and trade protection on the environment of various countries from the viewpoint of investment-related CO₂ emissions, which has important reference value for global CO₂ emission reduction in the context of the evolving trade situation.

1. Introduction

The Paris Accord was ratified in 2015, which states that the entire world must work together to keep the average global temperature rise below 2 °C [1]. Global warming is mostly caused by greenhouse gases, of which CO₂ emissions account for approximately 80% [2,3,4]. If people want to have a two-thirds probability of controlling global warming to 1.5 °C, the total CO₂ emissions in the future will be about 420 Gt [5]. The 2022 carbon budget report shows that global human activities are expected to emit 40.6 billion tons in 2022, ocean carbon sinks and land carbon sinks are expected to absorb 57% of global carbon emissions, and the remaining carbon budget is about 380 billion tons, which means that there is a 50% chance that the global average temperature will rise by 1.5 °C in the next 9 years. Existing studies have mainly explored the impact of different trade scenarios on CO₂ emissions from human activities. Some studies assume no trade protection measures. Cherniwchan studied the impact of NAFTA on pollution and found that trade liberalization contributes about two-thirds of the reduction in American manufacturing air pollution [6]. Shapiro and Walker found that trade liberalization increases productivity, which helps reduce CO₂ emissions. Some studies explored the impact of adopting trade protection measures on CO₂ emissions [7]. Liu et al. focused on the products and industries involved in the anti-dumping of 10 countries against China from 1995 to 2011, and found that the number of dumping cases sued and the CO₂ emissions of the industries involved in the anti-dumping cases showed a consistent trend of change [8]. Lu et al. explored the Sino-US trade conflict in 2018 and found that it led to land use changes and increased emissions in some developing countries [9]. In the context of economic globalization, many developed countries have promoted their own economic development through transnational investment, which increased CO₂ emissions outside the territory. China is the main investment target of developed countries. The increase in investment intensity has led to an increase in CO₂ emissions in China [10]. Investors move pollution-intensive industries to developing countries to reduce local investment-related CO₂ emissions. In terms of sorting out countries’ responsibility for CO₂ emissions, the effect of transnational investment on CO₂ emissions cannot be ignored.

In recent years, trade protectionism has gradually risen, and many studies have begun to focus on the effect of trade protection on CO₂ emissions. Tian et al. found that the removal of tariffs among RCEP members boosts CO₂ emissions from fuel combustion by about 3.1% over the same period [11]. Lin et al. found that increasing tariffs is beneficial to reduce global CO₂ emissions by 6.3% [12]. Some countries are trying to protect the popularity of local products. Draconian trade barriers are put in place. Du et al. found that trade restriction cannot make the environment and economy of the two countries have a win–win consequence through the Sino-US trade conflict [13]. Britain reduced trade with EU partners, but the positive effect of transnational trade on CO₂ emissions was not reduced [14]. Andersson studied the effect of China’s institutional reforms on CO₂ emissions from 1995 to 2008. After China joined the World Trade Organization in 2001, trade liberalization led to a rapid increase in China’s CO₂ emissions. The research results show that once the initial effect of trade liberalization fades, China’s growth in embodied emissions from trade with developed countries will slow [15]. Long et al. studied the impact of the 2008 economic crisis and the 2011 Tohoku Earthquake on Japan’s CO₂ emissions, and found that the economic crisis and natural disasters shut down the Japanese economy and reduced trade with other countries. This may have reduced Japan’s CO₂ emissions [16]. In 2020, COVID-19 emerged in full force. Countries greatly reduced the need for import and export trade in order to reduce the spread of the virus. Globally, international M&A fell by 15%, and the number of transnational project finance deals fell by a quarter [17]. Ray et al. investigated the effect of COVID-19 lockdown measures on global annual CO₂ emissions, focusing on 47 countries and their 105 cities that were severely affected by December 2020. Overall, the total carbon emissions of 184 select countries reduced by 438 Mt in 2020 compared to 2019 [18]. Trade protection measures have reduced the scale of transnational investment. This may indirectly reduce the implied investment-related CO₂ emissions. Numerous studies have examined the relationship between transnational investment and CO₂ emissions. Some have argued that transnational investment increased host-country pollution emissions, as developed countries transferred polluting industries to developing countries, sacrificing the environment of developing countries for their own economic growth. Nejati and Taleghani used a multi-regional general equilibrium model and found that if FDI does not have technology diffusion, it will increase the CO₂ emission intensity of the country [19]. Xu et al. found that long-distance trade is more conducive to the achievement of sustainable goals in developed countries than in developing countries [20]. Others hold the opposite view that developed countries spread their national advantages to developing countries through transnational investment, which improves energy efficiency in developing countries. Kisswani and Zaitouni found that the transfer of clean technology through FDI inflows to host countries reduced CO₂ emissions in Malaysia and Singapore [21]. Zmami and Ben-Salha found that in the short run, FDI flows are associated with less environmental pollution [22]. In order to provide a reference for policy makers to better grasp the environmental costs behind the evolution of the current trade situation and timely adjust investment policies according to possible changes in the future, this study quantifies the effect of trade protection on investment-related CO₂ emissions.

Based on the multi-regional input–output model (MRIO), we simulate investment-related CO₂ emissions of 16 global economies in both trade freedom and trade restriction. This study divides 43 countries and ROW into 16 economies. The specific economy divisions are provided in

Table A1. List and classification of economies.

Acronym	Economy
AUS	Australia
BRA	Brazil
CAN	Canada
CHE	Switzerland
CHN	China
IDN	Indonesia
IND	India
JPN	Japan
KOR	South Korea
MEX	Mexico
NOR	Norway
RUS	Russia
TUR	Turkey
USA	America
EU-28	Austria, Belgium, Britain, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, and Sweden (as of 2014)

of the Appendix A. This paper uses the undifferentiated technology assumptions to replace the trade restriction scenario with the no-trade scenario. The international trade is split into three trade links, namely the trade of final goods, the trade of end-stage intermediary goods, and the trade of remaining-stage intermediary goods. It enables the identification of countries’ positions in the global industrial chain and their effects on CO₂ emissions. This paper also explains the main drivers of the trade effect on investment-related CO₂ emissions in different countries from a sectoral point of view.

Here are the remaining sections of this study. Section 2 reviews the relationship between investment and CO₂ emissions and presents the paper innovation points. Section 3 demonstrates the study model and databases of the paper. Section 4 discusses the effect of investment on CO₂ emissions and its results from the viewpoint of the country, sectors, and trade links. Section 5 summarizes the results and points out the shortcomings.

2. Literature Review

Many developed countries transfer the environmental pollution generated by their own consumption demand to developing countries through international trade, transnational investment, and technology transfer, which results in global environmental pollution. Wiedmann and Lenzen found that the environmental pollution related to the export products of developed countries is transferred to developing countries in order to alleviate the pressure of their own environmental protection policies [23]. Zwab et al. found that international trade promotes the development of the global economy. But international trade emits more greenhouse gases than NTS [24]. Essandoh et al. studied the long-term correlation between FDI and CO₂ emissions, and found that in developed countries, FDI reduces CO₂ emissions, and in developing countries, FDI can promote CO₂ emissions [25]. The relationship between transnational investment activities and the environment has received a lot of attention. Numerous studies have been conducted around the effect of investment activities on the environment. Bu et al. found that multinational investment enterprises are more energy-efficient than domestic enterprises, taking the Jiangsu Province of China from 2005 to 2007 as the research object [26]. Zhang and Zhou adopted the STIRPAT model and found that FDI can help China reduce CO₂ emissions [27]. Sun et al. considered the long-term nature of FDI and found that a 1% increase in FDI will increase CO₂ emissions by 0.058% [28]. On the one hand, many studies have discussed the effect of domestic investment activities on CO₂ emissions. Cardarso et al. studied increased investment in Spanish tourism and found that Spanish tourism investment increases its carbon footprint responsibility based on a life cycle assessment model [29]. Ganda conducted a study on the effect of national investment in innovation and technology on CO₂ emissions [30]. Yang et al. analyzed how investments in renewable energy affect CO₂ emissions, and found that increasing investment in renewable energy reduces CO₂ emissions through technical effects, but increases CO₂ emissions through multiplier effects, and the structural effect of investment scale is not significant [31]. On the other hand, the effect of transnational investment on CO₂ emissions has received wide attention [32,33]. Zhang et al. studied the carbon footprint of multinational companies and found that the total amount of carbon transfers related to investment peaked in 2011 due to the decline in carbon intensity [34]. Shahbaz et al. studied how FDI affects CO₂ emissions, and found that CO₂ emissions in France are reduced due to the introduction of environment-friendly technologies by FDI [35]. Huang et al. found that rising levels of FDI and international trade are the root causes of environmental pollution [36]. Zhu et al. constructed the input–output decomposition framework and found that the scale effect makes FDI produce more CO₂ emissions, but the carbon intensity effect makes FDI reduce CO₂ emissions [37]. Ji et al. conclude that increased investment in China will promote economic growth and energy consumption at the same time based on the input–output model [38]. Most studies show the effect of transnational investment on CO₂ emission. This paper will indicate the effect of investment activities on CO₂ emissions from the viewpoints of countries, sectors, and trade links in order to more thoroughly evaluate the link between investment and CO₂ emissions.

Recently, trade protection has gradually emerged [39,40]. Nowadays, the environmental effect of trade barriers is widely studied. de Melo and Solleder conclude that by reducing tariff and non-tariff barriers, developing countries can benefit from better markets for environmental goods and services [41]. Liu et al. found that US–China trade friction helps reduce the effect on global CO₂ emissions [42]. Most studies focus on the environmental effects of trade liberalization and trade policies. Wang et al. studied the effect of trade on CO₂ emissions and found that when trade opening breaks through a certain breakpoint, CO₂ emissions will be suppressed [43]. Zhang et al. conclude that CO₂ emissions are inversely linked with exports and significantly linked with imports, taking 52 countries along the Belt and Road as research objects [44]. Wang and Wang found that in the long run, trade openness has a depressing effect on world CO₂ emissions [45]. Alola discovered that American trade policies increase CO₂ emissions in short order [46]. Wang and Zhang came to the conclusion that free trade significantly affects the dissociation of economic growth from CO₂ emissions in developed countries but inversely affects it in developing countries [47]. Wang and Wang found that with the increase in FDI, trade openness is obviously beneficial to the reduction in investment-related CO₂ emissions [48]. However, this literature does not specifically examine the effect of global trade protection measures on investment-related CO₂ emissions.

The environmental effect of production fragmentation is a focus of academic research. Liu et al. found that China produces more local CO₂ than other trading partners after joining the WTO [49]. With the effect of global manufacturing debris, countries are globally linked through different trades according to their technical and manufacturing characteristics. Each country’s position in the value chain is inextricably linked to the division of global responsibility for reducing emissions. Nowadays, more studies focus on the environmental effect of the links in international trade. Feng studied the effect of three trade patterns on CO₂ emissions and found that the trade of remaining-stage intermediary goods will contribute to the increase in CO₂ emissions [50]. For the division of trade links, Wang et al. decomposed the trade area triggered by domestic demand into three links, namely the trade of final goods, the trade of end-stage intermediary goods, and the trade of remaining-stage intermediary goods, in accordance with the Leontief inverse matrix’s decomposition findings [51]. Zhang et al. used the same decomposition method to divide trade into three links in order to study the effect of different trade links on global CO₂ emissions [52]. This paper follows this decomposition and divides international trade into three international trade links. On this basis, the CO₂ emission-accounting model of trade links in both normal and no-trade scenarios was constructed to research the effect of free trade on investment-driven CO₂ emissions of each link.

Three methods are widely used to study the effect of trade and investment on CO₂ emissions, which are an econometric analysis [53,54], a decomposition analysis [55,56,57], and the input–output method [58,59,60]. Wang and Zhang used the fully modified ordinary least squares method to find that the increase in R&D investment contributes to the decoupling of environmental pressure and economic growth [61]. Li and Li conclude that energy-related investments increase CO₂ emissions over all of China by constructing a spatial econometric model [62]. Andreoni and Galmarini used exponential decomposition to analyze the driving factors of CO₂ produced by energy in 33 countries, and found that economic growth is the main driving factor [63]. Wen et al. investigated the factors that affect CO₂ emissions by using a decomposition analysis and discovered that the scale of investment is the primary factor that increased CO₂ emissions [64]. Deng and Xu used the MRIO model and SDA to study the global embodied carbon trade from 1995 to 2009 based on WIOD [65]. Meng et al. analyzed the embodied CO₂ emissions of the South–South trade using a multi-regional input–output analysis, the bilateral trade-implied emissions method, and a structural decomposition analysis, and the results show that the embodied carbon emissions of the South–South trade increased by more than double during 2004–2011 [66]. Fu et al. investigated how the investment-driven economic growth affects a country’s CO₂ emissions using the input–output model [67]. Bi et al. coupled a multi-regional input–output model with an emission inventory to assess the impact of international trade on China’s atmosphere. The results show that in 2007, 50–60% of China’s air pollution emissions were affected by trade [68]. Wang et al. calculated the consumption of renewable and non-renewable energy in the global supply chain based on the multi-regional input–output model (MRIO) and a data envelopment analysis (DEA) [69]. Hong et al. assessed land-use emissions reflected in global trade from 2004 to 2017 based on a MRIO model and concluded that about three-quarters of the embodied emissions came from land-use change, with a significant portion transferred from low-income countries to highly industrialized areas [70]. One-third of the energy consumed and CO₂ released in China in 2007 was attributed to domestic investment, primarily from the manufacturing and construction industries. Compared with the other two analysis methods, an input–output analysis is more comprehensive and can save more time and manpower. This study uses a multi-scenario analysis approach to compare investment-driven CO₂ emissions in both trade freedom and trade restriction. Applying undifferentiated technology assumptions to the input–output model allows for quantitative simulation of no trade. Namely, when used, manufacturing technology of the importer is assumed to be identical to that of the home country, which is consistent with the fact that in the absence of trade, the demand for imported goods needs to be met by domestic technology manufacturing. MRIO tables are formed by linking inputs to demand for all sectors and all countries [71]. The effect of transnational investment on global CO₂ emissions is examined in this work using a MRIO model.

In general, this paper has two contributions. Firstly, most of the existing studies have targeted the effect of investment activities on CO₂ emissions, but few have addressed the effect of trade protection on investment-related CO₂ emissions. In order to make up for a gap in the literature, this paper examines how global trade protection measures affect the CO₂ emissions of transnational investment. This paper constructs an accounting framework for investment-related CO₂ emissions in both normal trade and no trade based on the MRIO model for the first time. This framework subdivides multiple trade links and trade sectors. Secondly, this study provides a reference for policy makers to better grasp the environmental costs behind the evolution of the current trade situation. And timely adjustment of investment policies according to possible changes in the future will play an important role in reducing investment-related CO₂ emissions.

3. Methodology

3.1. Calculation of Investment Trade: MRIO Model

A MRIO analysis is a quantitative analytical method for describing trade flows around the world, namely the economic trade between the sectors of the country where the inputs come from and the sectors of the country where they are distributed. It includes the single-regional input–output model (SRIO) and MRIO model. The SRIO model assumes that the rest of the world except the target country is a whole. It cannot analyze the trade between countries. The MRIO model reflects the role of countries in the global value chain and takes into account the differences in national technology levels. In this paper, the MRIO model is used to analyze the effect of transnational investment on global CO₂ emissions.

Assuming that the world consists of N economies, each of which consists of K sectors, the MRIO model is as follows:

$$\begin{array}{c}\left[\begin{array}{c}{X}_1\\ X_2\\ \dots \\ X_n\end{array}\right]=\left[\begin{array}{cccc}{A}_{11}& A_{12}& \cdots & A_{1n}\\ A_{21}& A_{22}& \cdots & A_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ A_{n1}& A_{n2}& \cdots & A_{nn}\end{array}\right]\left[\begin{array}{c}{X}_1\\ X_2\\ \dots \\ X_n\end{array}\right]+\left[\begin{array}{c}{Y}_1\\ Y_2\\ \dots \\ Y_n\end{array}\right]\end{array}$$

(1)

Here, $X_n$ typifies the gross output matrix of economy n, and $A_{nm}$ represents the intermediary consumption of products produced in economy n by unit output of economy m. The input coefficient matrix meets $A_{nm}^{sr}=Z_{nm}^{sr}/X_m^r$, including $Z_{nm}^{sr}$ (i, j = 1…n) that typifies the transfer from the sector s of economy n to the sector r of economy m. The intermediary input matrix from economy n to economy m is shown by $Z_{nm}=A_{nm}{X}_m$. $Y_n$ is the final input matrix for economy n. They are both economic variables in dollars.

$$\begin{array}{c}\left[\begin{array}{c}{X}_1\\ X_2\\ \dots \\ X_n\end{array}\right]=\left[\begin{array}{cccc}{I-A}_{11}& A_{12}& \cdots & A_{1n}\\ A_{21}& {I-A}_{22}& \cdots & A_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ A_{n1}& A_{n2}& \cdots & I-A_{nn}\end{array}\right]\left[\begin{array}{c}{Y}_1\\ Y_2\\ \dots \\ Y_n\end{array}\right]\end{array}$$

(2)

$$\begin{array}{c}\left[\begin{array}{c}{X}_1\\ X_2\\ \dots \\ X_n\end{array}\right]=\left[\begin{array}{cccc}{L}_{11}& L_{12}& \cdots & L_{1n}\\ L_{21}& L_{22}& \cdots & L_{2n}\\ ⋮& ⋮& \ddots & ⋮\\ L_{n1}& L_{n2}& \cdots & L_{nn}\end{array}\right]\left[\begin{array}{c}{X}_1\\ X_2\\ \dots \\ X_n\end{array}\right]+\left[\begin{array}{c}{Y}_1\\ Y_2\\ \dots \\ Y_n\end{array}\right]\end{array}$$

(3)

where $L_{nm}$, the Leontief inverse matrix, typifies the total input consumed by economy n in the production process of a unit of the final output of economy m. This coefficient relates final consumption to the resulting total output.

$$\begin{array}{c}{E}_{nm}=Y_{nm}+A_{nm}{E}_n\end{array}$$

(4)

$E_{nm}$ typifies the investment output of economy n to economy m, including the final investment demand $Y_{nm}$, and the intermediary investment demand $A_{nm}{E}_n$. $B_{nn}={\left(1-L_{nn}\right)}^{-1}$ typifies the native Leontief inverse matrix, $L_{nn}=B_{nn}+B_{nn}{\displaystyle {\sum }_{n\ne t}^i}{A}_{nt}{L}_{tn}$ [72].

$IE{f}_n$ is defined as the transnational investment demand of final goods, where investment partners directly absorb transnational investment products. $IE{i}_n$ is the transnational investment demand of end-stage intermediary goods, which needs further reproduction by the investment partner before it can be fully absorbed by the investor. $IE{g}_n$ is narrowly defined as the transnational investment demand of remaining-stage intermediary goods. Investment in the first two forms of trade crosses borders once and is eventually absorbed by the investors. Investment related to remaining-stage intermediary goods crosses more than once and requires direct or indirect investment between sectors in countries other than the home country.

$$\begin{array}{c}{E}_{nm}=\underset{\underset{IE{f}_n}{\leftrightarrow }}{Y_{nm}}+\underset{\underset{IE{i}_n}{\leftrightarrow }}{A_{nm}{B}_{mm}{Y}_{mm}}+\underset{\underset{IE{g}_n}{\leftrightarrow }}{A_{nm}{B}_{mm}{\displaystyle \sum _{t\ne m}^G}{A}_{mt}{B}_{tm}{Y}_{mm}+A_{nm}{\displaystyle \sum _{t\ne m}^G}{L}_{tm}{Y}_{tm}+A_{nm}{\displaystyle \sum _{t\ne m}^G}{L}_{mt}{\displaystyle \sum _{u\ne m}^G}{Y}_{tu}}\end{array}$$

(5)

We assume that the normal trade scenario is free trade and replace the trade protection scenario with the no-trade scenario based on the assumption of undifferentiated technology.

$$\begin{array}{c}{E}_n=A_{nn}{E}_n+Y_{nn}+{\displaystyle \sum _{n\ne m}^G}{E}_{nm}\end{array}$$

(6)

$$\begin{array}{c}{E}_n=B_{nn}\left(Y_{nn}+{\displaystyle \sum _{n\ne m}^G}{E}_{nm}\right)\end{array}$$

(7)

In Formula (6), $E_n$ typifies the total output used to meet domestic investment demand and transnational investment demand. Formula (6) can be replaced by Formula (7) according to $B_{nn}={\left(1-L_{nn}\right)}^{-1}$.

The following Formulas (8)–(12) typify the investment output formula under the scenario of normal trade. The total output of the total investment in economy n consists of four parts. $I_n$ typifies the output of domestic investment activities. The last three parts correspond to three trade patterns under the scenario of normal trade.

$$\begin{array}{c}{E}_n=I_n+E{f}_n+E{i}_n+E{g}_n\end{array}$$

(8)

$$\begin{array}{c}{I}_n{=B}_{nn}{Y}_{nn}\end{array}$$

(9)

$$\begin{array}{c}E{f}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{f}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}{Y}_{nm}\end{array}$$

(10)

$$\begin{array}{c}E{i}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{i}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}{A}_{nm}{B}_{mm}{Y}_{mm}\end{array}$$

(11)

$$\begin{array}{c}E{g}_n={B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{g}_n=B}_{nn}{\displaystyle \sum _{n\ne m}^G}(A_{nm}{B}_{mm}{\displaystyle \sum _{t\ne m}^G}{A}_{mt}{B}_{tm}{Y}_{mm}+A_{nm}{\displaystyle \sum _{t\ne m}^G}{L}_{mt}{Y}_{tm}+A_{nm}{\displaystyle \sum _{t\ne m}^G}{L}_{mt}{\displaystyle \sum _{u\ne m}^G}{Y}_{tu})\end{array}$$

(12)

In the absence of trade, transnational investment is prohibited. This means domestic technology will be used to produce products for domestic investment to meet domestic development needs. The output formula of each link under the scenario of no trade is as follows:

$$\begin{array}{c}{E}_n^{\prime }=I_n+NE{f}_n+NE{i}_n+NE{g}_n\end{array}$$

(13)

$$\begin{array}{c}{I}_n=B_{nn}{Y}_{nn}\end{array}$$

(14)

$$\begin{array}{c}NE{f}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{f}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}{Y}_{mn}\end{array}$$

(15)

$$\begin{array}{c}NE{i}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{i}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}{A}_{mn}{B}_{nn}{Y}_{nn}\end{array}$$

(16)

$$\begin{array}{c}NE{g}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}IE{g}_n=B_{nn}{\displaystyle \sum _{n\ne m}^G}(A_{mn}{B}_{nn}{\displaystyle \sum _{t\ne n}^G}{A}_{nt}{B}_{tn}{Y}_{nn}+A_{mn}{\displaystyle \sum _{t\ne n}^G}{L}_{nt}{Y}_{tn}+A_{mn}{\displaystyle \sum _{t\ne n}^G}{L}_{nt}{\displaystyle \sum _{u\ne n}^G}{Y}_{tu})\end{array}$$

(17)

3.2. Calculation of CO₂ Emission

This section describes the calculation process of CO₂ emission.

$$\begin{array}{c}{f}_n^s=\frac{G_n^s}{E_n^s}\end{array}$$

(18)

$f_n^s$ typifies the CO₂ emission intensity of sector s of economy n, $G_n^s$ typifies the CO₂ emission of sector s of economy n, and $E_n^s$ typifies the total output of sector s of economy n. The matrix F is a diagonal matrix consisting of $f_n^s$. Then, the carbon intensity matrix F of the world is

$$\begin{array}{c}F=\left[\begin{array}{cccc}{F}_1& 0& \cdots & 0\\ 0& F_2& \cdots & 0\\ ⋮& ⋮& \ddots & 0\\ 0& 0& 0& F_n\end{array}\right]\end{array}$$

(19)

Based on Formulas (8) and (19), investment-related CO₂ emission of economy s under the scenario of normal trade can be written as

$$\begin{array}{c}{G}_n=F_n{I}_n+F_{n}E{f}_n+F_{n}E{i}_n+F_{n}E{g}_n\end{array}$$

(20)

$F_n{I}_n$ typifies the CO₂ emissions produced by domestic investment activities of economy n under the scenario of normal trade, $F_{n}E{f}_n$ typifies the transnational investment-related CO₂ emissions of final goods, and $F_{n}E{i}_n$ typifies investment-related CO₂ emissions of end-stage intermediary goods. $F_{n}E{g}_n$ typifies investment-related CO₂ emissions of remaining-stage intermediary goods.

Based on this, according to Formulas (13) and (19), it can be concluded that investment-related CO₂ emission of economy s under the scenario of no trade is

$$\begin{array}{c}{G}_n^{\prime }=F_n{I}_n+F_{n}NE{f}_n+F_{n}NE{i}_n+F_{n}NE{g}_n\end{array}$$

(21)

$F_n{I}_n$ is the same as the result of CO₂ emission generated by domestic investment under the scenario of normal trade. $F_{n}NE{f}_n$, $F_{n}NE{i}_n$, and $F_{n}NE{g}_n$, respectively, typify investment-related CO₂ emissions caused by the trades of final goods, end-stage intermediary goods, and remaining-stage intermediary goods in the case that economy n does not import and completely depends on domestic technology manufacturing.

Based on both normal trade and no-trade scenarios, the effect of investment on CO₂ emissions in global trade links is studied in this research.

$$\begin{array}{c}∆G=G_n-G_n^{\prime }\end{array}$$

(22)

$∆G$ indicates the difference between investment-related CO₂ emissions in both normal trade and no-trade scenarios, $∆f_n=F_{n}E{f}_n-F_{n}NE{f}_n$ indicates the trade effect on investment-related CO₂ emissions of final goods, $∆i_n=F_{n}E{i}_n-F_{n}NE{i}_n$ indicates the trade effect on CO₂ emission of end-stage intermediary goods, and $∆g_n=F_{n}E{g}_n-F_{n}NE{g}_n$ typifies the trade effect on investment-related CO₂ emission of remaining-stage intermediary goods.

3.3. Data Source

This paper uses two data sources: CO₂ emissions and MRIO tables. Currently, multi-regional input–output databases include EORA, GTAP, WIOD, etc. Although EORA contains the complete time series from 1990 to 2021, it only divides 26 sectors. The GTAP database cannot provide a complete time series, which is very unfavorable for studying the trade rules of continuous years. After comparison and weighing, this study used the WIOD database. The MRIO tables from WIOD provide researchers with detailed data on global structural change and economic growth [73,74]. The 2016 MRIO table provides time series data at current prices from 2000 to 2014. This period is representative of trade development and includes some important events in international trade, such as China’s 2001 WTO membership and the 2008 global financial crisis. The table includes 43 individual countries as well as other areas (ROW) that encompass all countries that have not been listed separately. Currency data for each country are derived from 56 different sectors. The countries in the MRIO table are consolidated into 15 individual economies, the EU-28, and ROW, and the sectors are consolidated into seven sectors for simplicity of calculation. We used the sum of total fixed asset formation (in millions of national currencies) and changes in inventories and other valuables (in millions of national currencies) in the final demand of the MRIO table as actual investment capital. To eliminate the effect of inflation, all currency data within that paper were converted into constant 2010 dollars. The WIOD’s Environmental Accounting Database was utilized to obtain CO₂ emission data by section and country from 2000 to 2014 to calculate the carbon intensity.

4. Empirical Results

4.1. The Overall Effect of Global Trade on Investment-Related CO₂ Emissions

In the context of economy globalization and free trade, investment-related CO₂ emissions continued to rise from 2000 to 2014. They increased by 68.3% from 802.135 million tons in 2000 to 1353.465 million tons in 2014 (Figure 1). The effect of the global financial crisis on international trade resulted in a 2.04% drop in 2009. China, the US, India, Russia, and Japan contributed a large proportion in the global investment-related CO₂ emissions. From 2000 to 2014, China’s economic growth and export growth were fast, and its investment-related CO₂ emissions increased by 236%, followed by India’s 132%. This may be because China joined the WTO in 2001 and accelerated international trade cooperation. As two major developing countries, China and India have developed rapidly in recent years, which has strengthened international trade. In contrast, investment-related CO₂ emissions in Australia and the EU-28 showed a downward trend, falling by 19.8% and 19%, respectively. This result also confirms the current situation of international embodied carbon transfer to a certain extent. By transferring emissions-intensive and low value-added industries to developing countries, developed countries have not only gained economic benefits but also reduced their own environmental pressure.

Investment-related CO₂ emissions are caused by purely domestic investment activities and transnational investment activities. From 2000 to 2014, global investment-related CO₂ emissions under the scenario of no trade were more than those under the scenario of normal trade, and they peaked in 2011, with the largest difference of 546.17 million tons between them.

Figure 2 shows investment-related CO₂ emissions of 16 economics in both normal and no-trade scenarios in 2014. Investment-related CO₂ emissions of most countries under the scenario of no trade are higher than those under the scenario of normal trade. China, the US, India, Russia, and Japan are the major contributors to the world’s CO₂ emissions, whose total investment-related CO₂ emissions take into account 67.28% of the general investment-related CO₂ emissions around the world in 2014. Investment-related CO₂ emissions of India, the US, and China under the scenario of no trade are greater than those under the scenario of normal trade, and investment-related CO₂ emissions of Russia and Japan under the scenario of no trade are smaller than those under the scenario of normal trade. Among them, India’s investment-related CO₂ emissions under the scenario of no trade increased the most, by 18.86% compared with those under the scenario of normal trade. Investment-related CO₂ emissions under the scenario of no trade are the most reduced than those generated by normal trade in Russia. That is a decline of 12.70%. This indicates that the transnational investment reduction is beneficial to the decrease in CO₂ emissions in Russia. Russia has become a “pollution refuge” for developed economies such as the US. Developed economies transfer high-pollution and low-value investment industries to labor-intensive Russia with a low threshold of low-carbon policies. Furthermore, the speedy expansion of international trade cooperation increases Russia’s investment-related CO₂ emissions. Normal trade has boosted capital flows in Russia and Japan, but it has also led to more CO₂ emissions. In addition to the above countries, investment-related CO₂ emissions in the EU-28, South Korea, Mexico, and Norway under the scenario of normal trade are greater than those under the scenario of no trade. Trade protection is beneficial for these countries to lower investment-related CO₂ emissions.

4.2. Trade Pattern’s Effect on Investment-Related CO₂ Emissions

For each trade link in Figure 3, it consists of two parts: trends in investment-related CO₂ emissions and the difference between investment-related CO₂ emissions under normal trade and no-trade scenarios. Figure 3 shows the trade of end-stage intermediary goods contributed for the biggest amount of investment-related CO₂ emissions among three trade links. Its share increased from 38.52% in 2000 to 42.19% at the peak point in 2014. This is a reflection of the swift expansion of end-stage intermediary goods trade during this period. CO₂ emissions from the trade of remaining-stage intermediary goods remained relatively stable. But the global financial crisis in 2008 had an effect on international trade, making the share of CO₂ emissions from remaining-stage intermediary goods trade fluctuate greatly in 2009, reaching the lowest point of 24.26%. The portion of investment-related CO₂ emissions from final goods trade remained above 35% before 2009, peaked at 38.60% in 2009, and then continued to decline to the minimum of 31.81% in 2014. In the absence of trade, the trade of end-stage intermediary goods is still the main link of investment-related CO₂ emission contribution. The portion of investment-related CO₂ emissions in this link’s overall CO₂ emissions remained above 40% after 2006 and increased to 47.53% in 2013. CO₂ emissions from remaining-stage intermediary goods trade were still stable.

This paper divides the effect of trade protection on investment-related CO₂ emissions into three trade links. Figure 4 shows the change trend of the difference in investment-related CO₂ emissions between normal trade and no trade in each trade link of the five economies. From the viewpoint of three trade links, CO₂ emission of the US and India mainly comes from the trade of final goods and the trade of end-stage intermediary goods while CO₂ emission of China mainly comes from the trade of end-stage intermediary goods and the trade of remaining-stage intermediary goods. For the EU, free trade is conducive to the reduction in investment-related CO₂ emissions in the trade of remaining-stage intermediary goods. At the same time, it promotes the increase in investment-related CO₂ emissions in the trade of final goods and the trade of end-stage intermediary goods. In Russia, investment-related CO₂ emissions from three trade links under the scenario of normal trade were all greater than those generated under the scenario of no trade before 2003. Since 2004, investment-related CO₂ emissions from final goods trade under the scenario of normal trade are less than those generated under the scenario of no trade. In 2012, the difference reached a maximum of 78.05 million tons, which indicates that the investment under the scenario of normal trade contributes to a decrease in CO₂ emissions in Russia’s final goods trade. The opposite results of investment-related CO₂ emissions generated by three trade links show that different trade links affect various countries’ transnational investment-related CO₂ emissions in a variety of ways, which is closely related to the country’s investment policy, manufacturing technology, and capital structure.

4.3. Sectoral Effect on Investment-Related CO₂ Emissions

This section can be seen as a deeper investigation of the sectoral drivers that led to the findings in the first section, which examined investment-related CO₂ emissions from the viewpoint of trade links. Figure 5 shows the sectoral contribution to global investment-related CO₂ emissions, mainly concentrated in mining, construction and power, gas, and water supply sectors. The specific sectoral divisions of S1, S2, and S7 are provided in

Table A2. Classification of sectors.

Code	Integrated Sector	Sector in WIOD
S1	Agriculture	Crop and animal production, hunting, and related service activities
		Forestry and logging
		Fishing and aquaculture
S2	Mining	Mining and quarrying
S3	Manufacturing	Manufacturing of food products, beverages, and tobacco products
		Manufacturing of textiles, wearing apparel, and leather products
		Manufacturing of wood and products of wood and cork, except furniture; manufacturing of articles of straw and plaiting materials
		Manufacturing of paper and paper products
		Printing and reproduction of recorded media
		Manufacturing of coke and refined petroleum products
		Manufacturing of chemicals and chemical products
		Manufacturing of basic pharmaceutical products and pharmaceutical preparations
		Manufacturing of rubber and plastic products
		Manufacturing of other non-metallic mineral products
		Manufacturing of basic metals
		Manufacturing of fabricated metal products, except machinery and equipment
		Manufacturing of computer, electronic, and optical products
		Manufacturing of electrical equipment
		Manufacturing of machinery and equipment n.e.c.
		Manufacturing of motor vehicles, trailers, and semi-trailers
		Manufacturing of other transport equipment
		Manufacturing of furniture; other manufacturing
		Repair and installation of machinery and equipment
		Publishing activities
S4	Electricity, Gas, and Water Supply	Electricity, gas, steam, and air conditioning supply
		Water collection, treatment, and supply
S5	Construction	Construction
S6	Transport	Land transport and transport via pipelines
		Water transport
		Air transport
		Warehousing and support activities for transportation
S7	Commercial and public services	Sewerage; waste collection, treatment, and disposal activities; materials recovery; remediation activities; and other waste management services
		Wholesale and retail trade and repair of motor vehicles and motorcycles
		Wholesale trade, except of motor vehicles and motorcycles
		Retail trade, except of motor vehicles and motorcycles
		Postal and courier activities
		Accommodation and food service activities
		Motion picture, video, and television program production; sound recording and music publishing activities; and programming and broadcasting activities
		Telecommunications
		Computer programming, consultancy, and related activities; information service activities
		Financial service activities, except insurance and pension funding
		Insurance, reinsurance, and pension funding, except compulsory social security
		Activities auxiliary to financial services and insurance activities
		Real estate activities
		Legal and accounting activities, activities of head offices, and management consultancy activities
		Architectural and engineering activities; technical testing and analysis
		Scientific research and development
		Advertising and market research
		Other professional, scientific, and technical activities; veterinary activities
		Administrative and support service activities
		Public administration and defense; compulsory social security
		Education
		Human health and social work activities
		Other service activities
		Activities of households as employers; undifferentiated goods- and services-producing activities of households for own use
		Activities of extraterritorial organizations and bodies

of the Appendix A. For example, in 2014, these three sectors’ respective investment-related CO₂ emissions were 4362.94 million tons, 4246.74 million tons, and 2756.38 million tons, respectively, making up 85.66% of the global total investment-related CO₂ emissions. Next came business and public services, followed by agriculture sectors, manufacturing sectors, and transportation sectors. Resulting from the effect of the global financial crisis, the development of all sectors slowed down from 2008 to 2009, and investment-related CO₂ emissions all decreased. Among them, agriculture sectors and commercial and public service sectors had the greatest effect, with a decrease of 9.79% and 7.77%, respectively. The three sectors of agriculture, manufacturing, and electricity, natural gas, and water supply increased their CO₂ emissions by 177.65%, 127.35%, and 131.12%, respectively, in 2014 compared with 2000. They increased the most among all the sectors and were the most unfriendly to the environment. When compared to other sectors, the construction sector only experienced a 13.47% increase, and investment-related CO₂ emissions grew at a moderate rate.

Figure 6 displays the investment-related CO₂ emissions distribution of various countries in both normal and no-trade scenarios according to the analysis in 2014. Figure 6 represents that mining sectors, electricity, natural gas, and water supply sectors, construction sectors, and business and public service sectors contribute a lot in various countries. Of all the countries, China generates the highest amount of investment-related CO₂ emissions in every sector, which also confirms that China has the largest investment-related CO₂ emissions. Among them, investment-related CO₂ emissions from agriculture sectors, mining sectors, and electricity, natural gas, and water supply sectors—which produce greater investment-related emissions—account for 60.51%, 52.29%, and 57.50% of the world’s overall CO₂ emissions, respectively. The sectors in which India invests the most CO₂ emissions are also agriculture, mining, and electricity, natural gas, and water supply sectors, which to some extent indicates that China and India have the same major development industries in terms of international investment. Through the comparison of the two figures, we found that different national sectors’ investment-related CO₂ emissions are severely affected by trade protection.

Figure 7 shows the distinction between the two scenarios of no trade and normal trade, where the positive value typifies the positive-environmental effect of trade and the negative value typifies the negative-environmental effect of trade. It can be seen from Figure 7 that international trade has different effects on investment-related CO₂ emissions in different sectors of countries. In China, the CO₂ emissions generated by investment under the scenario of normal trade are less than those under the scenario of no trade in agriculture sectors, manufacturing sectors, and electricity, natural gas, and water supply sectors, decreasing by 7.39%, 24.72%, and 3%, respectively. Under the scenario of normal trade, the environment is negatively affected by the mining sector, while the trade situation has no significant effect on investment-related CO₂ emissions of China’s construction sectors, transportation sectors, and commerce and public service sectors. Seven sectors in Russia generate more investment-related CO₂ emissions under the scenario of normal trade than those under the scenario of no trade. Among them, the manufacturing sectors have the greatest effect. Investment-related CO₂ emissions under the scenario of normal trade are 6.4 times more than those generated by investment under the scenario of no trade. CO₂ emissions are negatively affected by transnational investment in Russia, which needs to further optimize the investment trade structure. In the EU-28 and Japan, normal trade has a detrimental effect on CO₂ emissions from mining and the electricity, gas, and water supply sectors, while having a favorable effect on CO₂ emissions from manufacturing. In India, normal trade has a significant effect on investment-related CO₂ emissions from manufacturing sectors, mining sectors, and power, gas, and water supply sectors. International trade reduces investment-related CO₂ emissions in the US from mining sectors and the electricity, gas, and water supply sectors, but increases investment-related CO₂ emissions from business and public services.

5. Conclusions

This paper establishes an accounting framework for investment-related CO₂ emissions under trade freedom and trade restriction based on the MRIO model for the first time, and studies the effect of global trade protection on investment-related CO₂ emissions in the three levels of the country, section, and trade links. In this paper, trade activities are divided into three trade links: the trade of final goods, the trade of end-stage intermediary goods, and the trade of remaining-stage intermediary goods. The paper examines investment-related CO₂ emissions at each stage of international trade links as well as the effect of two trade scenarios on CO₂ emissions.

According to this study, transnational investment-related CO₂ emissions increased by 68.3% in 2014 compared with 2000. And China, the US, India, Russia, and Japan contribute significantly more to the global investment-related CO₂ emissions. Global investment-related CO₂ emissions between 2000 and 2014 are higher under the scenario of no trade than those under the scenario of normal trade, with a maximum difference of 546.17 million tons. Normal trade is more conducive to reducing global investment-related CO₂ emissions. From the viewpoint of countries, trade protection is quite disadvantageous to CO₂ emissions reduction in China and India, and their CO₂ emissions will respectively increase by 105 million tons and 141.5 million tons compared to normal trade. Trade protection can decrease investment-related CO₂ emissions of Russia and Japan. From the sectoral viewpoint, the electricity, gas, and water supply sectors and the manufacturing sector are the main sectors for investment-related CO₂ emissions. From the viewpoint of trade links, the most significant portion of emissions is attributable to the trade of end-stage intermediary goods under the scenario of normal trade, which peaked at 42.19% in 2014. Trade protection can increase investment-related CO₂ emissions in the trade of end-stage intermediary goods. In addition, the trade scenarios have different effects on investment-related CO₂ emissions of each trade link in each country. Normal trade helps reduce investment-related CO₂ emissions in the trade of final goods and the trade of end-stage intermediary goods in the US, but it increases investment-related CO₂ emissions in the trade of remaining-stage intermediary goods. Normal trade helps increase investment-related CO₂ emissions in the trade of end-stage intermediary goods and the trade of remaining-stage intermediary goods in Russia.

The research results provide a reference for measures to reduce investment-related CO₂ emissions, such as promoting trade openness and reducing trade barriers to achieve a reduction in global total investment-related CO₂ emissions. China and India can attract transnational investment to reduce investment-related CO₂ emissions in both countries. If China focuses on introducing clean technologies for transnational investments in manufacturing, this will help reduce investment-related CO₂ emissions from manufacturing. Strengthening China’s barriers in the trade of final goods or focusing on the energy efficiency of this trade link will have a favorable effect on reducing investment-related CO₂ emissions in this trade link. For Russia and Japan, it is beneficial to pay attention to investment quality and optimize the investment structure to reduce investment-related CO₂ emissions. In general, trade liberalization has a negative effect on investment-related CO₂ emissions from the trade of end-stage intermediary goods. The trade barriers to investment related to intermediate products in the trade of end-stage intermediary goods on a global scale should be reduced, so as to achieve the greatest possible reduction in investment-related CO₂ emissions in this link. There are still expansions worth exploring in this study. Firstly, this paper uses a domestic technology hypothesis to study investment-related CO₂ emissions under the scenario of no trade. However, it is a fact that not every economy possesses the necessary resources or technology to accomplish this. Secondly, this research applies the MRIO table from 2000 to 2014 in the world input–output database, which has a certain lag for now. With the update of the MRIO table, the results also need to be updated.