2.1 EE-MRIO Model

The environmentally extended multi-regional input-output (EE-MRIO) model can analyze multiple environmental impacts driven by production or consumption and describe the input-output relationship between sectors and regions in a specific area or multiple sectors in different areas (Boulay, Hoekstra, & Vionnet, 2013; Meng et al., 2018; Zhao & Chen, 2014). For a multi-input-output table with m regions and n sectors, there is a value balance that can be expressed as a matrix on the equation relationship:

where \(X\) is the total output of all regions with \(mn \times 1\) dimension, \(A\) is a complete matrix composed of the direct input coefficient matrix of each region with \(mn \times mn\) dimension, \(Y\) represents the final demand matrix with \(mn \times 1\) dimension, each element \(y^{rs}\) is a \(n \times 1\) vector, which can be further divided into household consumption \(y_{household}^{rs}\), government consumption \(y_{government}^{rs}\), and investment \(y_{investment}^{rs}\) in the framework of the GTAP10 database, representing the three types of final demand from region r to region s.

Equation 1 can be further rearranged as:

\(I\) is an identity matrix with \(mn \times mn\) dimension, \(L\) is an inverse matrix of Leontief with \(mn \times mn\) dimension.

2.2 Selection of relevant sectors

To measure the environmental impact at the sectoral level, specific sectoral scopes need to be further defined. Due to research needs, we aggregated the 65 sectors in GTAP10 into 45 sectors based on sector similarity. Briefly, we have made certain modifications to the original codes in the GTAP10 database, and the corresponding sector code and sector description are shown in Table 1. The research scope of this study is the independent 45 sectors in new sector numbers and codes.

2.3 Data source and processing

The available GTAP database is balanced and has enough information to construct a multiregional input-output table. Therefore, we have combined various data to form various sub-matrices of the multiregional input-output table for 2014 by referring to Peters, Andrew, and Lennox (2011). The process was as follows: first, we used the GTAPAgg2 program, which is used to aggregate GTAP sectors, to aggregate the original 65 GTAP sectors to the 45 sectors needed for the study; then, we extracted the relevant data with the help of the ViewHAR program (a GTAP database visualization software) and Excel; finally, we distributed and adjusted the import data and the domestic data for intermediate use and final consumption, obtaining a multi-regional input-output table of 141×45. In addition, since bilateral trade data from region r to region s only provide information on export product sector \(i\) and not on consumption sector \(j\), as described in Peters et al. (2011), we allocated bilateral exports (in the prices of the producing country) proportionally to the intermediate use sector and final consumption according to the import structure of the importing region (in prices of the consuming country), which ensures output balance while constructing the non-diagonal matrix \(Z^{rs}\) and \(y^{rs}\).

\(Z^{rs}_{ij}\) represents the input from sector \(i\) in region \(r\) to sector \(j\) in region \(s\), \(y^{rs}_{i}\) represents the final consumption of region \(s\) for the products of sector \(i\) in region \(r\), \(vim^{s}_{i}\) represents total import of \(i\) into region \(s\), \(e^{rs}_{i}\) represents export of \(i\) from region \(r\) to region \(s\), \((Z^m)^{s}_{ij}\)represents the use of products in sector \(i\) by sector \(j\) in region \(s\), \((y^m)^{s}_{i}\) represents the final consumption of products in sector \(i\) by region \(s\).

The energy consumption of each sector in the 141 countries and regions is taken from the corresponding GTAP10 database account. Since the GTAP database does not contain sector-level freshwater withdrawals for 2014, we extrapolated global sector-level withdrawals through multiple calculation methods. Based on the definition of water use (National Bureau of Statistics [NBS], 2015), we divided water withdrawals into three categories (i.e., agricultural, industrial, and residential withdrawals). The freshwater withdrawals data and the proportions of the three types of withdrawals were obtained from the World Development Indicators (WDI) database (World Bank, 2021) . The three types of freshwater withdrawals were assigned to 45 sectors. The proportions of water use in the agricultural subsector by country in 2014 were obtained based on freshwater withdrawals in the agricultural subsector from the 2011 GTAP water accounts. Assuming that the relative water withdrawal intensity between the industrial subsectors is constant for each country and does not significantly vary across the different regions, the ratios of industrial subsectors were used to allocate the total industrial amount into subsectors and obtained from Fan, Wang, Zhang, Kong, and Song (2019). The distribution of water withdrawal in the service sector is similar to the industrial sector, while ecological water withdrawal was not considered in this study. For the regions with missing data in total water withdrawal or industrial distribution, sectoral water intensity was replaced by the global average water intensity at the sectoral level.

3.1 The overview of energy and water footprint of East Asian countries

The total global energy use embodied in the products of production (consumption) and exports (imports) in 2014 were 9,649 and 7,573 Mtoe, respectively. There is a significant difference between the embodied use of energy for production-based accounting (PBA) and consumption-based accounting (CBA) in East Asian countries (Figure 1). CBA embodied energy use is greater than PBA embodied energy use in Japan, Mongolia, and other East Asian countries or regions. And, PBA embodied energy use is only 51 Mtoe higher than CBA embodied energy use in South Korea. In contrast, China’s PBA embodied energy use (2,687 Mtoe) is significantly higher than its CBA embodied energy use (2,295 Mtoe), suggesting that China has undertaken many embodied energy exports to meet the final demands in the rest of the world. At the same time, closely related to the large population, China’s production-side and consumption-side embodied energy use both have a large global share, accounting for 28% and 24%, respectively. In addition, China’s embodied energy of exports and imports account for only 13% and 26% of the total on the consumption and production sides, indicating that most of the energy supply is self-sufficient in China’s domestic market, while the proportions of energy imports and exports in embodied energy use for the rest East Asian countries are all at a high level.

In terms of the different end-use categories, in East Asia, household consumption is dominant in the end-use of embodied energy for all countries except China. Specifically, the share of household consumption in Japan, South Korea, Mongolia, and other East Asian countries is above 60%, while the share of household consumption in China is 45%. In addition, energy use is driven by investment consumption and accounts for a relatively large proportion of the energy use structure of various countries, whereas the embodied energy use on the production side has similar results.

In 2014, the global total water use for production (consumption) and export (import) reached 398 billion m³ and 332 billion m³, respectively. The embodied water of CBA/imports is greater than PBA/exports for all countries except for other regions in East Asia where there is less embodied water, indicating that East Asia is a net importer of embodied water in global trade relations (Figure 2). China’s PBA embodied water and CBA embodied water are large, accounting for over 10% of the world total, with 52,671 and 59,181 million m³ of water used on the production and consumption side. Japan and South Korea have a relatively small world share of PBA and CBA embodied water, with CBA embodied water higher than PBA embodied water. Similar to energy use, imports and exports are still the main reason for embodied water use on the consumption and production side in all countries except China, where exports and imports account for only 26% and 17% of the consumption and production side.

The embodied water driven by household consumption in East Asia accounts for a large proportion of all embodied water. The proportion of embodied water use of residents in five countries or regions is more than 60%. Among them, the embodied water use of residents in Japan and Mongolia accounts for more than 80%, and the total amount is 98.13Gm³ and 0.52Gm^3, respectively. China has a relatively low water use share of 70% for household consumption, but due to its extremely large population, its total household water use is higher at 69.49 Gm³.

3.2 The sector structure of embodied energy and embodied water in East Asia

There is significant heterogeneity in embodied energy use driven by household consumption among different sectors in both the production and consumption sides in East Asia (Figure 3). In terms of embodied energy use on the production side, industry sectors, like the electricity, petroleum and coal products sectors, are the main energy-using sectors in East Asian countries. In South Korea and Japan, PBA energy use in both sectors are over 90%, while PBA energy use in the electricity sector of Mongolia accounts for 88% of the total production-side embodied energy use. In terms of CBA embodied energy use on the consumption side. At the same time, industry sectors, dominated by the petroleum and coal products sectors are still the main sectors consuming energy in each country in East Asia. The services sectors have increased significantly in terms of their share in end-use energy, becoming the most energy-intensive sectors in East Asian countries in addition to the industry sectors. The share of embodied energy use in the service sectors of the total in South Korea, Japan, and Mongolia is 44%, 41%, and 28%, respectively.

Since embodied energy use driven by investment consumption in China is another important energy use component, it was further analyzed. The results show that the findings of the PBA and CBA embodied energy use driven by household consumption in China are similar to those of other East Asian countries, with the electricity sector and the petroleum and coal products sector accounting for 83% of the total in the PBA perspective, and the services sectors accounting for 26 % in the CBA perspective. An analysis of investment consumption shows that the electricity sector and the petroleum and coal products sector are the main end-use sectors in the PBA perspective, with a combined share of 74% of the total. However, in the CBA perspective, the construction sector and heavy manufacturing sector are the main energy-use sectors, with a share of 56% and 23%, respectively, indicating that these sectors not only attract a large amount of fixed capital inflows but also drive a large amount of energy use.

Figure 4 illustrates the sectoral structure of embodied water use driven by household consumption in East Asian countries. The agriculture and processed foods sectors are the most important water-use sectors on the production and consumption side. The sectoral distribution of embodied water use in East Asia from the CBA perspective shows diversity and complexity compared to those from the PBA perspective. On the production side, the agricultural sectors, especially the paddy rice sector, are the key water-use sector in the major East Asian countries, except for the rest of East Asia, which is associated with high irrigation water demand for cultivation. For example, the agricultural sector embodied water accounts for 89.4% of Japan’s total water. In comparison, the paddy rice sector alone accounts for 84.1% of the total water footprint. The share of agricultural sectors’ water use in South Korea and China is 89% and 83%, respectively, with the paddy rice sector embodied water accounting for 78% and 35% of the total, respectively. Mongolia’s share of embodied water in agriculture sectors is 65%, and unlike the other countries, the vegetable, fruit and nut sector accounts for the largest share of embodied water at 63%. Moreover, there is significant heterogeneity in the sectoral distribution of embodied water use between the consumption and production sides. Specifically, processed foods, industry, and services sectors are key water-use sectors in most countries. The share of embodied water in processed food sectors is over 40% in China, Japan, and South Korea, and the proportion of the processed foods sectors’ embodied water accounts for nearly 60% of the total in China and Japan. The processed rice sector in the processed foods sector is also the main contributor to the total amount of embodied water, accounting for 31% and 21% of total embodied water in Japan and South Korea. The share of embodied water in the industry sectors and service sectors ranges from 10% to 20% in China, Japan and South Korea. Related to its dietary preferences, the sectoral structure of Mongolia’s embodied water differs significantly from that of China, Japan, and South Korea, mainly in its higher water use of primary agricultural products on the consumption side.