Towards Environmental Sustainability: An Input–Output Analysis to Measure Industry-Level Carbon Dioxide Emissions in Egypt

Abstract

Egypt’s average share of global carbon dioxide emissions has been rising from mid-1990s to date. This motivates the present study to identify industries that drive carbon dioxide emissions (as direct emitters and as total emitters with high emission multiplier effects). Environmental input–output analysis is applied to Egypt’s 2017–2018 input–output table to measure sectoral emissions. The industries identified as high emitters are linked to Egypt’s achievement of Sustainable Development Goals, namely, Goals 7, 8, 9, 12, and 13. The findings indicate that ten industries qualify as environmentally degrading (dirty), having the highest emission multiplier effects (in descending order): electricity, gas, and water; non-metallic mineral products; basic metals; rubber and plastic products; chemicals and chemical products; paper and paper products; food products; hotels and restaurants; transportation and storage; and textiles. Eight of these industries also have high output multiplier effects. This underscores that although potential investment in and the growth of these industries will generate output multiplier effects, they will also be coupled with emission multiplier effects. Five other industries had high emission multipliers, as follows: water and sewerage; beverages; coke and refined petroleum products; extraction of crude petroleum; and mining of metal ores. The growth of these industries would not be in favor of the achievement of SDGs. Policy measures are recommended.

1. Introduction

The United Nations Framework Convention on Climate Change Conference (UNFCCC) continues to underscore the critical role of greenhouse gas (GHG) emissions (primarily carbon dioxide (CO₂), methane, and nitrous oxide) in trapping heat within the atmosphere. This has intensified the greenhouse effect, which drives global warming and climate change. Since 2015, when the Paris Agreement set the target of limiting the global temperature rise to 1.5 °C, the world has been on track to exceed this limit [1]. Against this backdrop, the UNFCCC calls on all sectors, businesses, and institutions to make solid commitments to reduce emissions. Much of the empirical literature on emissions thus seeks to identify businesses/sectors that act as drivers of CO₂ emissions, nominating the ones that must make commitments to reduce emissions.

The present study thereby undertakes the research problem of identifying the sectors with high direct CO₂ emissions as well as those with high total CO₂ emissions (generating direct plus indirect emissions) in Egypt. Direct emissions stem from the industry’s own output, while indirect emissions stem from its linkages with upstream industries along its supply chain and downstream industries utilizing its output for intermediate use. It is believed that such identification is a step towards reducing CO₂ emissions, recommending those sectors that should make commitments to CO₂ emission reduction, and calling on the government of Egypt to strengthen action to reduce emissions.

This study employs the environmental input–output (EOI) method of analysis to measure direct and total CO₂ emissions for Egypt’s industries. It relies on data from the most recent 2017–2018 national input–output (IO) transaction table. For a robustness check, the direct and total CO₂ emissions by industry are also computed using the preceding year’s IO table (2016–2017). As customary in IO analysis, the term “industry” is used interchangeably with the term “sector” or “activity” [2], (p. 10), denoting any economic activity within the IO table without distinguishing between ‘agricultural’, ‘industrial’, or ‘services’. The term “industry” is thus used in this study to refer to any economic activity within the IO table. The 2016–2017 and 2017–2018 IO tables were issued by Egypt’s Institute of National Planning (INP) and made available by the Central Agency for Public Mobilization and Statistics (CAPMAS). The detailed study results obtained using the 2016–2017 and 2017–2018 IO tables are given in Appendix B.

The interest in measuring emissions by industry in Egypt is itself driven by the country’s high (fossil fuel) CO₂ emissions, which totalled 209.5 million metric tons (mmt) in 2015, rising to 219.4 mmt in 2016, 243.29 mmt in 2020, and 265.96 mmt in 2022. Meanwhile, the country’s average share of global CO₂ emissions rose from 0.35% over the period of 1995–2005 to 0.58% in 2006–2016 and 0.64% in 2017–2022 (source: calculated from [3]).

This study further aims to explore how the use of EOI analysis and the identification of environment-degrading industries can be linked to Egypt’s achievement of the United Nation’s Sustainable Development Goals (SDGs), namely, Goals 7, 8, 9, 12, and 13 (discussed further in the literature review), and to corporate carbon disclosure. Both links have been established in the empirical emission literature. As for the former, EIO analysis provides empirical data that can inform policies aimed at sustainable production. Further, by quantifying the environmental consequences of the growth of industries, EIO can support evidence-based policymaking towards achieving the SDGs [4]. With regards to corporate carbon disclosure, the identified industries point to firms that need to consider carbon disclosure while also considering the effect of such disclosure on their financial performance and market competitiveness.

The present study is novel in being the first one to use EOI analysis to measure Egypt’s direct and total carbon emissions by industry. It is also novel in that it explores the relationship between the identification of these industries and Egypt’s achievement of the SDGs. To the best of the author’s knowledge, this study fills a gap in the empirical literature as neither EOI analysis for emissions measurement nor the exploration of the link to the SDGs has been previously conducted for Egypt.

The rest of the paper is organized as follows: Section 2 provides the research context, highlighting the conceptual framework and selected empirical studies of relevance; Section 3 describes the materials and methods, with a description of the data, and data sources, as well as a detailed description of the environmental input–output (EOI) analysis; Section 4 presents the results for industries’ computed output multipliers, the estimated direct CO₂ emissions, and the total emissions (the emission multipliers) in a scatter plot (with numerical values of multipliers using the two IO tables given in Appendix B). The results are discussed in Section 5, and Section 6 concludes and provides relevant policy implications.

2. Research Context

2.1. Conceptual Framework

IO analysis can be defined as a “top-down linear macroeconomic approach to describe industrial structure” [5]. While a detailed description of the analysis is given below, it is worth noting that this analysis continues to be widely used. This owes to its ability to unveil the level and nature of industrial interdependence in an economy. IO analysis further allows for measuring output multiplier effects as well as other multiplier effects, including employment and environmental effects. The output multiplier effects result from the direct output increase in an industry as well as the indirect output increases. The latter result from that industry’s reliance on the output of its upstream (supplying) industries and feeding of its own output to downstream (receiving) industries (for intermediate use). Moreover, as household income increases, spending on economic activities increases both directly and indirectly, thereby stimulating further output [6]. Additional induced effects thereby result from the overall increase in spending in the household sector.

A study [7] provided an extension of “the basic conceptual framework of the static input–output analysis” in the augmented original national IO tables [7] (p. 262). These tables included labor inputs (or row entries giving total employment by industry) and the amount of pollutants produced by industry, thus making it possible to obtain a complex structural matrix of national income [7] (p. 265). An analysis of the augmented IO table (as described below) would yield estimates of pollutant multiplier effects [6], thereby laying the foundation for EOI, which “reveals the channels through which the environmental burdens of production activities are transmitted throughout the economy” [8] (p. 3410).

It is also worth noting that some nations issue augmented IO tables that integrate environmental accounts with the nation’s basic IO table (as per the Leontief ‘augmented’ national IO tables). These are referred to as national income accounts, including environmental accounts (NAMEA). The NAMEA system therefore “encompasses environmental satellite accounts compatible with economic national accounts, meaning that economic flows, physical flows, and emissions can be linked together” [9] (p. 170). The typical Eurostat network NAMEA tables give row entries as the environmental pressure variables by industry and private households [10] (p. 71). The environmental pressure variables are broadly divided into two groups: residual input pressure and residual output pressure. Residual inputs include the primary energy supply, the total material requirement (of fossil fuels, metals, industrial and construction minerals, and biomass), and land use (built-up areas for settlement and traffic). Residual outputs include aggregated emissions of greenhouse gases (carbon dioxide, nitrous oxide, and methane), aggregated emissions of acidifying substances (sulfur dioxide, nitrogen oxides, and ammonia), and aggregated emissions of ozone precursor substances (total non-bulky and bulky (construction and demolition) waste generation).

NAMEA table column entries provide the total environmental pressure variables by environmental pollutant type. Thus, the table entries of environmental pollutants by industry can be denoted as g_ik (where i is the emitting industry and k is the environmental pressure variable). g_ik entries make it possible to obtain the environmental pressure coefficients (g_ki) as the emanated environmental pressure k variable relative to the output of the industry “i” (noting that the output value of the ith industry by row has the same output value as that of the same jth industry by column). Subsequently, it would be possible to obtain the environmental pressure inverse matrix (à la Leontief inverse) and the total direct and indirect emissions by industry, enabling an assessment of the direct and indirect emissions over the production chain [11] (p. 226).

The merits of using EOI analysis are threefold. First, EOI analysis can evaluate the interconnectedness of industries and quantify total (direct and indirect) emissions. It further helps identify environment-degrading industries. Second, EOI allows for linking these industries to a country’s performance toward achieving the SDGs and to corporate carbon disclosure. Third, IO analysis is among the important circular economy assessment methods, especially with its application to industries [12] (p. 447).

Regarding quantifying total emissions and identifying environment-degrading industries, EOI would help direct policymaking toward specific sectoral interventions. This would raise two important policy questions pertaining to the ‘high-production–high-emissions’ sectors: (1) Is the higher production multiplier effect jeopardizing environmental quality? (2) If so, what policy measures need to be put in place? EOI would thereby support governments in assessing the environmental sustainability of a country’s industrial practices. It would also allow them to assess whether their resource management is responsible in a way that fulfills the needs of current generations without compromising the needs of future generations.

A further advantage of identifying environment-degrading industries is to link them to a country’s performance toward achieving the SDGs. This is of special relevance to this study as it aims to specifically link the environment-degrading industries to Egypt’s achievement of Goals 7, 8, 9, 12, and 13 of the SDGs. Goal 7 aims to ensure access to affordable, reliable, sustainable, and modern energy for all. As energy intensity is an essential element in the computation of industry emissions, it is believed that these computations will inherently capture the relation of energy to achieving Goal 7. Of further relevance to achieving Goal 7 is a country’s use of renewable energy and its sustainable energy policies [13].

EIO analysis is also believed to incorporate the interaction of industries along their supply chains. Analysis results will thus lead to the promotion of sustainable and inclusive economic growth, i.e., Goal 8. It will similarly work towards achieving Goal 9 and Goal 12, which aim to build resilient infrastructure, promote inclusive and sustainable industrialization, and ensure sustainable production. As for Goal 13, which targets taking urgent action to combat climate change and its impacts, it is believed that the identification of environment-degrading industries may direct policymaking toward the necessary measures to combat climate change (for a full discussion of the applicability of IO analysis to tracking performance with regard to the SDGs, see [4]).

In the same vein as EIO analysis’ link to the achievement of SDGs comes corporate carbon emission disclosure. Many studies on emissions have established that corporate emission disclosure plays a role in sustainable industrialization and production. The literature discusses conditions under which firms may be compelled to disclose carbon emissions and how that enhances their financial performance (especially regarding lowering capital costs). These conditions include firms’ recognition that climate change can constitute damage to their business infrastructure and can entail additional costs of complying with climate change regulations. Moreover, firms can be subject to pressure from consumers and other stakeholders to take proactive action against climate change. Firms may also lose their competitive advantage, especially against competitors who are more responsive to consumers’ environmental concerns. Greener production and reduced carbon emissions of operations may minimize their resource use and lower costs, further motivating their carbon emission disclosure [14].

A study [15] identified stakeholder pressure (from shareholders, corporate headquarters, customers, suppliers, employees, financial institutions, environmental groups, and the media) as a key driver of carbon emission disclosure. Using an automated narrative analysis of the annual reports of a sample of UK firms from 2013–2019, the authors provided a measure of carbon emission disclosure which captures firms’ actual CO₂ emissions, as well as their environmental, social, and governance (ESG) scores.

The authors of [16] studied the effects of ESG reporting on the cost of capital (debt and equity) of UK non-financial firms from 2014 to 2018. These firms spanned sectors of technology, manufacturing, energy, retail, and health. The authors tested the hypothesis that ESG reporting has a negative effect on capital cost. By demonstrating a positive approach to environmental and social performance, as well as governance policies, firms may be able to reduce the information asymmetry between themselves and investors. Reduced capital costs may also be associated with lowering political costs associated with the negative perception of stakeholders about the firm’s environmental impacts [16] (p. 331).

2.2. Empirical Literature

Several empirical studies have employed EIO, and a few of them are reviewed to shed light on the wide scope of EIO use. A study [11] estimated the GHG emissions produced from Brazilian economic activities in 2009. The authors found the highest CO₂ emitters to be livestock; cement; other mining and quarrying; steel manufacturing; food and beverage manufacturing; and transport and storage services. Some authors [17] estimated the direct and indirect emissions from all industries in China in 2012 and found production-based emissions to be concentrated in the traditional energy- and carbon-intensive sectors, namely, the production and supply of electric and heat power, the smelting and pressing of ferrous metals, and the manufacture of non-metallic mineral products. These industries were reported to contribute 48.5%, 16.4%, and 7.6% of the total production emissions, respectively [17] (p. 269). Another study [18] found the petrochemical industry to rank highest in China’s total direct and indirect emissions (accounting for 18% of China’s total direct and indirect emissions), which mainly owes to the production and supply of electrical and heat power. Meanwhile, the following industries were ranked as having the next highest emissions: smelting and pressing of ferrous metals; manufacture of non-metallic mineral products; smelting and pressing of non-ferrous metals; transport, storage, and post services; and mining and washing of coal [18] (p. 1585).

Ref. [19] used NAMEA-type tables for Germany for 1999 to estimate total emissions. The authors found the following industries to rank as the highest GHG emitters: services of construction; manufacturing of food products and beverages; motor vehicles, trailers, and other transport equipment; basic metals; fabricated metal products; services of electricity, gas, and water; agricultural products, forestry, and logging; and services of retail trade, all of which fall along their respective production chains.

Ref. [5] employed EIO to assess the environmental impacts of a proposed airport in Sydney, Australia. They assessed the direct onsite emissions plus the total emissions to be 5.5 mt of GHG, compared to 0.24 mt of direct onsite GHG emissions. Thus, direct plus indirect emissions are almost 23 times the direct emissions alone (note: direct emissions were assessed by the Australian Bureau of Agricultural and Resource Economics) [5] (p. 272).

Furthermore, ref. [20] estimated the emission multipliers (particulates, sulfur dioxide, hydrocarbons, carbon oxide, and nitrogen oxide) for the region of Macedonia in 1998 and concluded that water and electricity services have the highest pollution multipliers. They were followed by the manufacturing of non-metallic mineral products, the manufacturing of fabricated metal products, construction services, oil products, agricultural products (primarily maize), the manufacturing of rubber and plastics, and the manufacturing of chemical products. The authors also estimated the industries’ output and employment multipliers. They indicated that there is always a conflicting relationship between industries that serve as targets of economic development for their employment and output considerations on the one hand and their degrading effect on the environment on the other hand.

In conclusion, the cited empirical studies shared many of the industries identified as high carbon dioxide emitters (both directly and in total). Furthermore, some of those industries were also identified by [21] as being high carbon dioxide-emitting industries: manufacturing of basic (ferrous and non-ferrous) metals; chemicals and chemical products; coke and refined petroleum products; non-metallic mineral products; rubber and plastics; fabricated metal products, except machinery; mining and quarrying of metal and non-metal ores; and extraction of crude petroleum.

Three of the other commonly identified high-polluting industries, which appeared in the cited works but not in [21], were the manufacturing of food products, beverages, motor vehicles, trailers, and other transport equipment. This suggests that the cited empirical works found them to be polluting because of their inter-industry linkages as opposed to their direct emission effects.

Thereby, the prospect of identifying polluting industries (through both direct and total emissions) is highly relevant to the case of Egypt. The reasons for selecting Egypt are threefold: First, there is a scarcity of studies that have conducted EOI analysis for Egypt. The present study thus aims to fill a gap in the empirical literature. Second, Egypt’s average share of global emissions has been steadily increasing from 1995 to date. For perspective, the respective average share of global emissions was 0.35% in 1995–2005, rising to 0.58% and 0.64% in 2006–2016 and 2017–2022, respectively. Such a steady rise in the average share of global emissions calls for a close examination of the emission-driving industries. Third, in 2017 (the issue date of the present IO table), industries’ contributions to total CO₂ emissions in Egypt were as follows: power industry: 32.1%, transport: 22%, and other industrial combustion: 18.7% (calculated from [3]). The respective industry’s contribution data again underscore the importance of identifying high-emission industries (direct and total) to help undertake relevant mitigation measures.

Furthermore, the reason for using the IO transaction table for the year 2017–2018 is that this issue is the latest available IO table for Egypt. While acknowledging that six fiscal years have elapsed since the table’s issue date, using this issue is justified on the grounds that multipliers are assumed to be stable during a typical period of up to six years after the initial calculation of the IO table [22]. Accordingly, this study uses the 2017–2018 issue under the confident assumption that the output and emission multipliers computed based on this table will provide an adequate reflection of the industries’ multiplier effects.

A study [15] indicated that since capital expenditure leaves a higher carbon footprint on the environment, firms with high capital expenditure tend to disclose more information on carbon emissions in their annual reports [15] (p. 112592). Firms are prompted to undertake such disclosure to avoid negative reactions in the market if that information is not disclosed. The findings further indicate that there is a positive relation between firms’ internal governance (including board size and independence, gender diversity, audit committee size, and independence) and carbon emission disclosure. This positive effect of governance is also underscored by [16], which found that the interaction of internal governance with the ESG score of firms negatively and significantly affects their capital cost. This implies that ESG reporting can lead to reduced capital cost, particularly in firms with robust corporate governance [16] (p. 337).

3. Materials and Methods

3.1. Data

Data for IO transactions were obtained from the ‘Input–Output Table for the year 2016/2017 at Basic Prices according to Economic Activities’ and the ‘Input–Output Table for the year 2017/2018 at Basic Prices according to Economic Activities’. The respective tables were issued by Egypt’s Institute of National Planning (INP) and the Central Agency for Public Mobilization and Statistics (CAPMAS). For details of the method employed in preparing the IO tables for Egypt, please see [23].

The main steps of data processing are summarized in Figure 1.

Data for the industry energy use value and industry output value are drawn from the 2017/2018 Egyptian Economic Census issued by the Central Agency for Public Mobilization and Statistics (CAPMAS).

Data for energy use for Egypt are drawn from the International Energy Agency [24].

Data for standard tons of CO₂ emitted per petajoule of different energy forms are drawn from [25].

3.2. Method

This section describes the static IO analysis and how it is employed to obtain the output multipliers. It further describes how IO analysis is extended to capture the environmental effects, thereby obtaining measures of the industry’s direct carbon emissions and total carbon emissions (direct and indirect emissions or the emission multiplier).

Static IO analysis describes and explains inter-industry transactions within the national economy. As it exposes the pattern of industry interdependence, IO analysis was used to investigate the structure of the national economy. This analysis enjoys the advantage of being transparent and having few assumptions built into it [26] (p. 439). These assumptions are as follows: (1) each industry produces one unique homogeneous product, and no two products are jointly produced by more than one industry; (2) prices, final demand, and factor supplies are given; (3) the supply of labor is perfectly elastic, such that adjustments in labor are quantity-based and are not related to any change in the wage rate; (4) there is free mobility of resources between industries; (5) the input–output relations between the industries are linear, such that a unit of change in the final demand for industry ‘j’ or a unit of change in the output of another industry that uses j’s production as an input will translate into a unit change in the production of the industry ‘j’; (6) the relationship between the industry’s inputs and outputs is fixed; thereby, the technical input–output coefficient (⍺_ij) is fixed (which is clearly an approximation to reality [27] (p. 272)); (7) the production in the IO system operates under constant returns to scale and ignores economies of scale in production and any possible substitution between inputs; (8) as industry ‘j’ buys from two industries (1 and 2), it uses the inputs from 1 and 2 in a fixed proportion, which is equal to the ratio of the two technical coefficients (such that p₁₂ = ⍺_1j/⍺_2j) (and with fixed ⍺_1j and ⍺_2j, p₁₂ is also fixed); (9) the technology of production is a fixed-proportion technology, which is also not affected by changes in the relative input prices of industry 1 and industry 2; and (10) technological progress is constant.

In its configuration, row entries in the IO table give the monetary value of sale transactions of the total output of industry “i” to all other industries “j” (i = j = 1,…, n). In turn, industry i is the producing industry that sells its output for (1) intermediate use (i.e., intermediate demand), denoted by X_ij, and (2) for final consumption (i.e., final demand), denoted by Y_i. Final consumption is final demand because the output goes to final use and is not an intermediate used in further industrial production processes. Final demand comprises purchases of the household sector, government sector, business sector (for private investment purposes), and foreign sector (sales of industry X_i abroad). Column entries in the table are those of the consuming industry j, namely, its intermediates purchased from all industries (including itself) to produce the output of industry j, and the value added, which comprises “wages, depreciation, profits, taxes, and other costs incurred in the industry” [7] (p. 265). The total output is the sum of the intermediates and the value added. The total of ‘i’ by row must equal the total output of ‘j’ by column since i = j.

X₁ = X₁₁ + X₁₂ + … + X_1n + Y₁

X₂ = X₂₁+ X_22v+ … + X_2n + Y₂

…

X_n = X_n1 + X_n2 + …+ X_nn + Y_n

(1)

Or:

$${\mathrm{X}}_i={\sum }_{j=1}^n{X}_{ij}+{\mathrm{Y}}_i, i = 1,...,n$$

(2)

IO analysis entails calculating the technical input–output coefficients denoted by a_ij’s, where each a_ij is the input of the selling industry “i” to the buying industry ‘j’ divided by the total output of industry ‘j’:

a_{ij =} X_ij/X_j

(3)

As such, the total output of industry X_i, i = 1, …, n, is given by

X₁ = a_{11 × 1} + a_{12 × 2} + … +a_1nX_n + Y₁

X₂ = a_{21 × 1} + a_{22 × 2} + … +a_2nX_n + Y₂

…

X_n = a_{n1 × 1} + a_{n2 × 2} + … +a_nnX_n + Y_n

(4)

Thus, matrix A of a_ij is defined as the technical input coefficient matrix (the Leontief matrix) for all industries, and matrix L = (I − A)⁻¹ is referred to as the Leontief inverse or the total requirements matrix. The elements of L, denoted by Ɩ_ij, measure how much of each industry i’s output is needed in direct and indirect requirements to produce one unit of industry j’s output (p. 10), where

X₁ = l₁₁ Y₁ + l ₁₂Y₂ + … + l _1nY_n

X₂ = l ₂₁Y₁ + l ₂₂Y₂ + … + l _2nY_n

…

X_n = l _n1Y₁ + l _n2Y₂ + … + l _nnY_n

(5)

Worth noting is that the increase in the gross output for every industry exceeds the increase in the final demand for that industry. This owes to industries being used in the production of other industries, making IO analysis “an investigation into the quantitative implications of such inter-industry relations” [27] (p. 272).

Using IO analysis, I obtain the industry output multiplier, which measures the total output from all domestic industries required to produce one additional unit of output of industry ‘j’ (the combined direct and indirect effects of a change in the final demand for the industry ‘j’). Multipliers are assumed to be stable during a typical period of six years after the initial calculation of the input–output table [22]. Therefore, computing output multipliers for Egypt using the 2017/2018 IO table would adequately reflect the present multiplier effects for Egypt’s industries. The output multiplier of industry ‘j’ is obtained from the Leontief inverse column sum as follows:

$${\mathrm{OM}}_j={\sum }_{i=1}^n{\mathrm{l}}_{ij}$$

(6)

As indicated in the above conceptual framework, some nations issue NAMEA-type tables, thus enabling EIO to be used to estimate emissions by industry. However, Egypt is among the nations that do not issue NAMEA-type tables. In case of the absence of NAMEA-type tables, it is possible to apply another method that extends the IO tables to include aspects of the economy–environment links [22] (p. 272). This study thus adopts the method of [22] to estimate (direct and indirect) carbon emissions by industry. The method is implemented in the following steps:

• Using data on the value of energy use by industry (in the form of oils; natural gas; coal; hydro; and biofuels) and the value of the industry’s final product, I obtain the ratio of energy use to the final product as a measure of energy intensity by industry.
• Using data for the nationwide energy use of Egypt in the two calendar years (2017 and 2018), I obtain a simple average to measure energy use for the fiscal year 2017/2018 (the year of the IO table).
• Energy use in Egypt comprises oil, natural gas, coal, hydro, and biofuels (measured in tetra joules). Accordingly, the share of each type of energy in total energy use is measured. This provides a measure of the energy mix for Egypt for 2017/2018: oil (37.8%); natural gas (54.5%); coal (2.7%); and hydro, biofuels, wind, and solar (5%).
• Using the standard tons of CO₂ emitted per petajoule (PJ) of each type of energy, (73,300 tons of CO₂/PJ from oil, 55,800 tons of CO₂/PJ from natural gas, and 93,900 tons of CO2/PJ coal), I obtain the level of CO₂ emitted from energy use with the given energy mix of Egypt (obtained in step 3). This is converted into millions of tons of CO₂ per tetra joule (TJ).
• The levels of CO₂ emitted from energy use (obtained in step 4) are multiplied by the energy intensity of each industry (obtained in step 1) to yield a measure of the direct level of CO₂ emitted for that specific industry based on its energy intensity and the energy mix of Egypt with the associated CO_2/TJ emissions.
• The calculated level of CO₂ emitted by industry (obtained in 5) is multiplied by the output multiplier calculated for each industry “j” (OM_j) to yield a measure of the direct and indirect levels of CO₂ emitted for industry “j” (i.e., total emissions for industry “j”).

In addition to overcoming the lack of publicly issued industry-specific CO₂ emissions data, the method employed in this study is appropriate for the economic context of Egypt, namely because it is data-driven. It captures the specific energy mix of Egypt with its associated CO₂ emissions level, as well as the energy intensity of each industry. This way, measures of industry-specific emissions are obtained. However, I do acknowledge that it may be limited in that it implicitly assumes that the energy mix of each industry mirrors the nationwide energy mix.

4. Results

This section presents results for the computed output multipliers and total (direct and indirect) CO₂ emissions (emission multipliers) of all industries, as shown in Figure 2. The results of the computed direct emissions, emission multipliers, and output multipliers based on the 2017/2018 IO table are shown in

Table A1. EOI analysis results for industries’ direct carbon emissions and total (direct and indirect carbon emissions) and output multipliers in millions of tons of CO₂ in 2016/2017 and 2017/2018.

Industry (ISIC Revision 4 Code)	Direct CO2 Emissions (Millions of Metric Tons)		Total (Direct and Indirect CO2 Emissions) Emission Multiplier in Millions of Metric Tons		Output Multiplier
	2016/2017	2017/2018	2016/2017	2017/2018	2016/2017	2017/2018
Agriculture, forestry, and fishing (01–03)
Crop and animal production and hunting; Fishing and aquaculture (01 + 02 + 03) (1)	1.121	0.2034	1.753	0.318	1.564	1.5628
Industry (Extraction and Mining) (06–09)
Extraction of crude petroleum (06)	1.716	1.5609	1.856	1.9077	1.082	1.2222
Mining of metal ores (07–09)	5.607	1.5380	7.873	1.8771	1.404	1.2205
Industry (Manufacturing) (10–33)
Food products (10) (1)	2.482	1.9573	5.238	4.3927	2.111	2.2443
Beverages (11)	1.166	1.1887	1.978	2.1860	1.696	1.8390
Tobacco products (12)	0.070	0.2110	0.109	0.3128	1.567	1.4827
Textiles (13)	1.809	1.6945	3.910	4.2383	2.161	2.5013
Wearing apparel (14)	0.829	0.7437	1.239	1.8005	1.494	2.4208
Leather and related products (15)	0.310	0.2000	0.598	0.2899	1.928	1.4491
Wood and cork, except furniture (16)	0.648	0.5374	0.723	0.9494	1.115	1.7665
Paper and paper products (17)	2.831	2.8515	5.654	6.4509	1.997	2.2623
Printing and reproduction of recorded media (18)	0.525	0.4088	0.573	0.8092	1.092	1.9794
Coke and refined petroleum (19)	1.549	1.0008	2.618	1.8804	1.690	1.8789
Chemicals and chemical products (20)	4.024	3.9003	6.944	7.8823	1.726	2.0209
Pharmaceuticals (21)	0.698	0.6996	1.310	1.4032	1.877	2.0057
Rubber and plastic products (22)	6.212	6.2580	12.076	13.2728	1.944	2.1210
Non-metallic mineral products (23)	16.024	15.8084	30.749	33.5035	1.919	2.1193
Basic metals (24)	6.746	6.6865	12.009	14.3954	1.780	2.1529
Fabricated metal products, except machinery (25)	0.728	0.6469	1.221	1.3247	1.677	2.0477
Computer, electronic and optical products (26)	0.101	0.1004	0.219	0.2404	2.172	2.3957
Electrical equipment (27)	0.419	0.4053	0.878	0.9576	2.098	2.3623
Machinery and equipment (28)	0.465	0.4527	0.954	0.9822	2.053	2.1697
Motor vehicles, trailers and semitrailers; Other transport equipment (29 + 30)	0.611	0.6018	1.325	1.4311	2.171	2.3781
Furniture (31)	0.311	0.2547	0.418	0.4997	1.3427	1.9616
Other manufacturing; Repair of machinery and equipment (32 + 33)	0.877	0.6400	0.897	0.9208	1.023	1.4389
Services (35–97)
Electricity and gas (35)	72.452	45.7810	126.911	108.8872	1.752	2.3784
Water and sewerage (36–38)	3.059	1.7761	4.967	3.1183	1.624	1.7557
Construction (41–43)	1.056	0.4689	1.640	0.9161	1.554	1.9537
Wholesale and retail; Repair of motor vehicles (45–47)	0.694	0.5586	0.835	0.6507	1.204	1.1649
Transport and storage (49–53)	14.110	3.1015	19.475	4.3707	1.380	1.4092
Hotels and restaurants (55 + 56)	2.590	2.5376	4.059	4.3784	1.568	1.7254
Communication (58 + 59 + 60 + 61 + 62 + 63)	0.840	0.3714	1.210	0.5465	1.440	1.4712
Financial services and insurance (64 + 65 + 66)	0.246	0.1386	0.279	0.1594	1.135	1.1497
Real estate (68)	0.441	0.0420	0.542	0.0453	1.228	1.0796
Professional services (69–75)	1.013	0.5513	1.299	0.7413	1.282	1.3446
Administrative and support services (77–82)	1.084	0.5618	1.662	0.7544	1.532	1.3429
Public administration and defense (84) (2)	NA	NA	NA	NA	1.3122	1.2948
Education (85)	1.031	0.2014	1.137	0.2641	1.103	1.3110
Human health activities; residential care; and social work activities (86–88)	0.227	0.0657	0.332	0.0988	1.462	1.5030
Arts and entertainment (90–93)	0.493	0.0944	0.601	0.1129	1.220	1.1964
Membership organization services (94–96)	0.411	0.2043	0.547	0.2723	1.331	1.3325

Notes: ⁽¹⁾ The 2016–2017 IO table includes only Crop and animal production, hunting, and related service activities (01) and Fishing and aquaculture (03). The 2017–2018 table includes Growing of cereals (except rice), leguminous crops, and oil seeds (01), Forestry and logging (02), and Fishing and aquaculture (03). ⁽²⁾ No direct or total emission multiplier values were obtained due to a lack of data for energy use of the sectors in the 2017–2018 economic census for Egypt. Source: author’s computations.

in Appendix B. Also shown in Table A1 in Appendix B are the direct emissions, emission multipliers, and output multipliers based on the 2016/2017 IO table computed as robustness checks for 2017/2018 computations. The ISIC Rev. 4 classification of industries in the 2016/2017 and 2017/2018 IO tables and the details of industry aggregation are also given in Appendix A.

Omitted from the industries shown in Figure 2 are electricity and gas, non-metallic mineral products, basic metals, rubber and plastic products, and chemicals and chemical products as they were classified as emission multiplier outliers. These industries’ emission multiplier values exceeded 7.81. Outlier values were calculated as quartile 3 + 1.5 × interquartile range, where the interquartile range = quartile 3 − quartile 1. The emission multiplier values greater than 7.81 were electricity and gas (108.9); non-metallic mineral products (33.5); basic metals (14.4); rubber and plastic products (13.3); and chemicals and chemical products (7.9); hence, they were all classified as outliers.

5. Discussion

Extracting key results from Figure 2, the findings indicate that:

• Ten industries (of which five did not appear in Figure 2 because they were emission outliers) were found to be the most environmentally-degrading (dirty) (arranged in descending order of emissions): electricity, gas, and water; non-metallic mineral products; basic metals; rubber and plastic products; chemicals and chemical products; paper and paper products; food products; hotels and restaurants; transportation and storage; and textiles.
• Eight of the above ten industries should be of particular concern to policymakers for possessing both high output and high emission multiplier effects. These industries are electricity, gas, and water; non-metallic mineral products; basic metals; rubber and plastic products; chemicals and chemical products; paper products; food products; and textiles.
• The following five industries are ranked as the next highest ones in terms of environmental concern, with emission multiplier effects greater than 1.9: water and sewerage; beverages; coke and refined petroleum products; extraction of crude petroleum; and mining of metal ores.
• Ten industries that qualify as environmentally non-degrading (clean) (in descending order of cleanliness) are (1) computers and electronic and optical equipment; (2) electrical equipment; (3) motor vehicles and trailers; (4) wearing apparel; (5) furniture; (6) printing; (7) machinery and equipment; (8) construction; (9) fabricated metal products; and (10) pharmaceuticals. All these industries have low emission multipliers coupled with high output multipliers (more than 1.8).
• The remaining industries are environmentally non-degrading, mostly service-related, and clustered in low-emission and low-to-intermediate output multiplier zones.

I note that the above results are largely in line with the findings of [11,17,18] and [19].

Linking the identified environmentally non-degrading (clean) and degrading (dirty) industries to SDGs 7, 8, 9, 12, and 13 involves understanding how the growth of these industries may contribute to or hinder the achievement of these goals. The findings are outlined in

Table 1. Industries identified as environmentally non-degrading (clean) and environment-degrading (dirty) and their relation to the achievement of SDGs 7, 8, 9, 12, and 13.

SDG	Industry
	Clean: Industry Growth May Contribute to Achieving the Goal	Dirty: Industry Growth May Hinder Achieving the Goal
Goal 7: ensure access to affordable, reliable, sustainable, and modern energy for all.	- computers and electronic and optical equipment; - electrical equipment; - motor vehicles and trailers; - wearing apparel; - furniture; - printing; - machinery and equipment; - construction; - fabricated metal products; - pharmaceuticals	- electricity, gas, and water; - non-metallic mineral products; - basic metals, - rubber and plastic products; - chemicals and chemical products; - paper and paper products; - food products; - hotels and restaurants; - transportation and storage; - textiles.
Goal 8: promotion of sustainable and inclusive economic growth.
Goal 9: build resilient infrastructure, promote inclusive and sustainable industrialization, and foster innovation.
Goal 12: ensure sustainable consumption and production patterns.
Goal 13: take urgent action to combat climate change and its impacts.

.

Furthermore, the detrimental effect of expanding the industries labelled as hindering the achievement of SDGs 7, 8, 9, 12, and 13 may be lessened if emissions mitigation measures are implemented. However, any relevant measures must be identified on an industry-by-industry basis. Similarly, firms in these industries need to implement carbon emission disclosure practices that are believed to lower their capital cost and enhance their competitiveness against environmentally friendly competition [14,15,16].

6. Conclusions

This study computed the direct and indirect carbon emissions (emission multiplier effects) generated from Egypt’s industries using EIO analysis based on the 2017/2018 IO table (the latest available for Egypt) while also computing the same multiplier effects based on the 2016/2017 IO table as a robustness check. The computed multipliers using the 2017/2018 IO table are assumed to be stable during a typical period of up to six years after the initial calculation of the IO table and are, therefore, expected to adequately reflect the current state of multiplier effects of Egypt’s industries.

The results obtained for industries’ direct and total CO₂ emissions (emission multiplier effects) show that the following industries rank as the highest direct carbon emitters: electricity, gas, and water; non-metallic mineral products; basic metals; rubber and plastic products; chemicals and chemical products; paper and paper products; food products; hotels and restaurants; transportation and storage; and textiles.

In addition, eight of these industries are of particular concern to policymakers (in terms of emissions) for having both high output and high emission multiplier effects (in descending order of importance): electricity, gas, and water; non-metallic mineral products; basic metals; rubber and plastic products; chemicals and chemical products; paper products; food products; and textiles. A key policy recommendation is that the growth of these industries must consciously weigh the output multiplier effects against the emission multiplier effects. Thereby, the following pressing policy questions should be answered: (1) if further output growth is going to come at the expense of increased emissions, what carbon dioxide mitigation and adaptation measures should be put in place? and (2) do we recognize that these measures must not only be undertaken in the industries themselves but also in the upstream and downstream industries along their value chains? All the proposed measures must be sector-specific, coupled with commitments to sector-specific CO₂ emission reduction.

Enabling and ensuring that all identified industries become energy efficient is a further policy priority. I hereby make special reference to the manufacturing industry of coke and refined petroleum, which was found to have an output multiplier of 1.8 as well as an emission multiplier of almost the same value. This industry has been identified in previous research as having suboptimal energy use [28] (p. 498). Thus, particular attention needs to be paid to the energy efficiency of the industry. In the same vein, reliance on renewable energy sources and the reduction in fossil energy are imperative.

Egypt also needs to progress in establishing a nuclear power station in the western desert since nuclear energy has been proven to be among the lowest CO₂ emitters [29], thus shifting from reliance on fossil-fuel-generated to nuclear-generated energy. It is further essential that Egypt adopts and applies the de-carbonization approach in the various stages of the value chains of industries, especially those identified as environmentally-degrading.

Further, the identification of industries whose growth may be a stepping stone towards the achievement of the SDGs 7, 8, 9, 12, and 13 is key from a sustainable development standpoint, and the present study has indicated that ‘preserving the environment’ and ‘growing’ need not be mutually exclusive targets. Policymaking must start by acknowledging that the high-output–high-emission-multiplier industries can indeed grow, but adequate mitigation and adaptation measures are needed for themselves and for industries along their value chains. Similarly, firms in these industries need to implement carbon emission disclosure practices, not only for transparency purposes but also because such disclosures are believed to lower firms’ capital costs and enhance their competitiveness.

Moreover, the placement of the top ten CO₂-emitting industries in industrial clusters (comprising, for example, the manufacturing of non-metallic mineral products, basic metals, and chemicals together) may be a step towards mitigating CO₂ emissions. The proximity of industries within clusters may help aggregate demand for energy across industries, creating opportunities for systemic efficiencies, electrification, demand optimization, and carbon capture by using the carbon in industrial and manufacturing processes and storing it underground where feasible [30]. Additionally, policymakers’ emphasis should be on the generation of green hydrogen by splitting water via electrolysis (using electricity from wind and solar) or by producing hydrogen from nuclear generation. Hydrogen is a promising technology for decarbonizing industries that are believed to be hard to abate. Further emphasis lies on the progress of Egypt’s nuclear power station in the western desert.

Of high policy relevance to the manufacturing industries, which make up nine out of the top ten emitters, is the optimization of internal processes to meet the growing demand for goods and services while minimizing environmental impacts [31] (p. 130). The firms of these industries should aim to adopt circular manufacturing (CM) strategies to extend products’ lifecycles, reduce resource use, and create a closed loop of resources. As a step in this direction, [32] proposed that firms begin by assessing the current state of not only their organization’s CM practices (related to products, processes, the management of external stakeholders, and tools such as technologies) but also of firms along their entire value chains [32] (p. 8). This should allow firms to “define a customized and structured improvement path” in their transition to CM [32] (p. 16).

Lastly, like any empirical work, this study has limitations. First and foremost, the EOI was conducted using the 2017/2018 IO table for Egypt, which was the latest one available. With a period of six years having elapsed since the issue of the IO table, one may be concerned about whether the multiplier results obtained reflect the current state of the economy. However, one advantage is that multipliers are assumed to be stable during a typical period of up to six years after the initial calculation of the IO table, suggesting that this study’s findings may be relevant.

Furthermore, while this study has attempted to describe the environmental impacts of all industries within the IO table, a detailed industry-specific carbon mitigation prescription is beyond its scope. It is advisable that industry-specific studies be conducted for the purpose of prescribing action plans for each industry. Future research may also build on this study by conducting a longitudinal study to monitor and track changes in emission multipliers after the emission reduction measures are implemented. Some firm-level survey quantitative approaches may be employed in industries identified as environmentally-degrading to provide insights into industries’ sustainability practices, such as carbon emission disclosure. This could help identify industry-level issues pertaining to sustainability practices and possible loopholes and recommend relevant policy actions.