Recently, the 21st United Nations Climate Conference (COP21) resulted in a worldwide agreement on climate pointing to the need to contain global temperature rise to under 2˚C, and, if possible under 1.5˚C. In general, the agreement considers a commitment of the parties to decrease emission levels based on their historic, current, and future responsibilities by establishing binding obligations in nationally determined contributions (NDCs), and to pursue domestic measures aimed toward achieving them. In addition, the agreement extended the current goal of mobilizing $100 billion a year in support by 2020 through 2025 with a new higher goal to be set for the period after 2025. In addition, COP21 called for a new mechanism, similar to the Clean Development Mechanism under the Kyoto Protocol, enabling emission reductions in one country to be counted toward another country’s NDC   .
According to Mexico’s intended NDC  , the country has committed to reduce 25% of its greenhouse gases (GHG) and short-lived climate pollutant emissions unconditionally (below “Business as Usual” scenario) by the year 2030. However, Mexico has a General Climate Change Law (GCCL) that establishes an aspirational objective of 30% reduction of emissions by 2020 and 50% by 2050 with respect to the emissions levels in 2000  .
There are several methods to evaluate energy consumption and GHG emissions and to identify mitigation opportunities for NDCs. Methods can be divided into bottom-up and top-down. Top-down models evaluate the system from aggregate economic variables, whereas bottom-up models consider techno- logical options or project-specific climate change mitigation policies in a model of energy systems  .
Input-output analysis is a top-down approach in which the data on production and consumption in all sectors allow a complete allocation of all activities to all products. GHG emissions are the result of economic activity that exists to meet human needs. Economic activity can be defined as all the production pro- cesses and the exchanges of goods and services between the productive sectors and the final demand. In that process, there is energy involved and therefore emissions. W. Leontief, a 1973 Nobel Prize winner, proposed input-output analysis  . The core of which is a table that shows the flow of goods and services, measured in monetary terms for a given time period, between the productive sectors that compose the economy and the final demand. It is a tool that allows analysis of the economy on a global scale and information of individual sectors at the same time. The main property of this technique is that it encodes the multiplicative effect  that comes from economic activity, allowing assessment of both direct and indirect effects.
Recent studies on energy consumption and GHG emissions using input?output analysis include: Alcantara and Padilla who developed input-output subsystems for the service sector in Spain that allowed the decomposition of the CO2 emissions into five different components: own, demand volume, feedback, internal, and spill-over components  ; Proops et al.  , who assessed the reduction of CO2 emissions in a comparative study for Germany and the United Kingdom; Tarancon et al.  , who used an input?output approach combined with a sensitivity analysis to analyze the direct and indirect consumption of electricity by 18 manufacturing sectors in 15 European countries. In addition, Tarancon and del Rio  provided a critical overview of sensitivity analyses within input-output techniques applied to energy-related CO2 emissions. Alcantara et al.  also analyzed the responsibility of the productive structure of an economic system with respect to the consumption and generation of electricity within an input- output framework.
Also, important studies using input-output models have been developed for China, the largest CO2 emission country, including three-scale input-output modeling for the urban economy  ; embodied energy, export policy adjustment, and China’s sustainable development: a multiregional input-output analysis  ; CO2 emissions of China’s food industry  ; urban carbon transformations: unraveling spatial and intersectorial linkages for key city industries based on multiregional analysis  ; and China’s regional disparities in energy consumption: an input-output analysis  .
More recently, input-output analysis has been used to estimate embodied emissions in trade. For example, Wiebe et al.  used input-output matrixes to analyze CO2 emissions embodied in international trade, covering 48 sectors in 53 countries and 2 regions. Su and Ang  analyzed emissions based on competitive and noncompetitive imports. Cortés Borda et al.  quantified the differences between production-based (territorial) and consumption-based (global) nuclear energy usage in the main 40 economies of the world through the application of a multiregional environmentally extended input-output model. Input- output matrixes are also the basis of the General Equilibrium Models applied for energy and CO2 emissions   that have gained importance for the analysis of climate policy impacts to the economy    .
In the case of Mexico, there are few analyses based on the input-output analysis related to GHG emissions. For example, Lewis  performed an input- output study of carbon dioxide emissions in Mexico linked to trade liberalization and the participation of Mexico in global trade  . Because of the lack of such analyses, this paper is novel in developing a top-down model based on input- output analysis for a middle-income country.
In this paper, we use input-output analysis to analyze the Mexican energy system and its GHG emissions for the year 2012 and to estimate a baseline scenario for 2026. The elasticity of emissions with respect to national demand is calculated in order to identify the total and distributed effects of CO2-eq emissions. In this framework, the analysis also evaluates the effects in the economy related to changes in individual sector demands and, vice versa, the effect on individual sectors due to global changes in national demands.
This paper is divided into four sections: Section 1 is the introduction, Section 2 presents the methodological framework as well as a brief description of the data used, Section 3 presents a discussion of the results, and Section 4 offers some conclusions.
2. Methodological Framework and Data Sources
According to input-output methodology, the economy can be decomposed on n sectors that produce and exchange goods or services. The bigger the number of sectors n, the more accurate and precise the model of the economy is. The input-output basic equation, also known as the Leontief equation is the following:
where is the total sectorial production, which is the sum of final demand and consumption among all sectors of economy, is the final sectorial demand, and is the Leontief matrix, where is the identity matrix and A is formed by that denotes the amount of product from sector i that is needed to produce one unit of product by sector j in monetary terms.
In order to estimate emissions, let n be the sectors of economy and K the number of different fossil fuel sources. Every sector is represented by vector and represents the amount of fuel k that sector i uses in one year. Let be the emission factor of fuel k and technology T. Then is the total carbon dioxide emissions (CO2-eq) of the economy related to fossil fuel combustion:
Let be the emission intensity, defined as the quantity of CO2-eq per unit of output of sector i. The vector is then the sectorial emission intensity formed by all (from). CO2-eq emissions of sector i will be the multiplication of the emission intensity of sector i by the activity of sector i:
. But substituting xi from Equation (1):
For the objective of this paper, the change in sectorial emissions due to the changes in final demand is then:
Let us define the elasticity of total CO2-eq emissions () due to changes in final demand of sector j as  .
From this point we can define a new variable: that takes into ac-
count the part j of total production that goes directly to final demand. This allows distinguishing between sectors whose production is mainly for satisfying final demand and those whose production is used as inputs by other sectors, therefore:
But considering (1), (6) can be expressed as the multiplication of two matrixes:
This matrix expression gives us the total emission variation of the economy due to a unitary change in final demand of all n sectors. To extract from here the emission variation of sector i, due to a change in final demand of sector j, we have to remove the sums from expression (6). Let be the diagonal matrix of vector and also
Then, the elasticity can be written as:
And for ij
The element represents the percentage of increase in CO2-eq emissions of sectors i in response to a 1% increase in final demand of sector j. For example, if sector i is agriculture and sector j is food industry, then would express the percentage of increase in CO2-eq emissions of agriculture in response to a 1% increase in final demand of the food industry. The matrix modifies the multipliers contained on Leontief’s matrix using emission intensities (emissions per monetary unit produced) that referred to the proportion of the total greenhouse gases emitted. Considering the definition of the elements, if we construct a column vector whose elements represent a percentage of increase on sectorial final demand, and multiply them by the matrix, then the result must be the sectorial percentage of increase in CO2-eq emissions in response to changes of all final demands:
where the 0 super index indicates the final demand in year 0, and 1 indicates the final demand at the end of an arbitrary time period. Let vector be multiplied by matrix then:
where is the i-th element of vector. The vector represents the percentage of increase in CO2-eq emissions in response to changes in the final demand of all sectors, assuming that the structure of the economy and emission intensity will remain constant. Equation (12) allows calculating base scenarios considering the variation in final demand.
2.2. Total, Distributive, and Structural Effects
From Equation (9), it is possible to analyze different effects that final demand has on emissions levels. The Total Effect (TE) is the sum over i (columns of the matrix), and represents the change in emissions for all the economy due to a unitary change in final demand of sector j. The Distributive Effect (DE) is the sum over j (rows of the matrix) and represents the change in emissions for all the economy due to a unitary change in each of the j sectors.
In addition, it is possible to separate each effect into two components  : Own Sector Effects (OE) that result from the changes in the final demand of each Own Sector (the diagonal elements of matrix), and Structure Effects (SE) that result from the changes in other sectors of the economy.
2.3. Data Sources
The input-output model constructed in this work comes from a combination of two different databases available in Mexico, in addition to the IPCC emission factors  . These are the 2012 input-output matrixes provided by the National Institute of Statistics and Geography (INEGI)  with three different levels of sectorial disaggregation (i.e., 19 sectors, 70 subsectors, and 262 branches) and the 2012 National Energy Balance (NEB)  with a sectorial disaggregation of 17 producing sectors in addition to the agricultural sector, commercial sector, and transport sector, which subdivides itself into 4 sectors. In total, the NEB provides 26 different sectors. In order to match energy sectors from NEB and economy sectors from the input-output matrix, some sectors were summed up either in energy or I-O matrixes (Table 1). CO2, methane (CH4), and nitrous oxide (N2O) emissions are considered. The CO2-eq for CH4 and N2O, are 21 and 298 respectively.
2.4. Final Demand Projection for Year 2026
Final demand is projected to 2026 using Equation (13). The annual rate of growth was projected from 2003 to 2014 to 2026 (3.5% per year). Fuel structure, economy structure, and CO2-eq intensity are considered constant.
3. Results and Analysis
Table 2 presents CO2-eq emissions related to Mexican energy consumption and production in 2012 and estimations of a baseline scenario for 2026, as well as the variation in percentage. Changes in final demand carry a total emission increase of 3.4%.
The total CO2-eq emissions impact matrix was calculated according to Equation (10) and is presented in Table A1. Table 3 presents the TE, and Table 4 the DE, both for 2012. The diagonal elements of both of Table 3 and Table 4 are the percentages of OE in TE and DE, respectively. A large TE (final column of Table 3) means that the sector’s final demand has a high influence on total emissions, whereas a large DE (final column of Table 4) means that an overall change in final demand has a large influence on emissions from the specific sector. For example, a 1% increase in final demand of the coal mining sector would lead to a 0.07% increase of total CO2-eq emissions (Table 3, row 1, final column), whereas
Table 1. Sectors for input-output analysis: sector code.
when there is a 1% increase of the final demand of all sectors, the emissions of coal mining would increase 0.16% with respect to the previous total emissions (Table 4, row 1, final column). The largest emissions come from the road transport sector. Therefore, a 1% increase in final demand of this sector would lead to a 333% increase of total CO2-eq emissions (Table 3, row 9, final column), and a 1% increase of the final demand of all sectors will represent an increase of 367% with respect to the previous total emissions (Table 4, row 9, final column).
Figure 1 presents the sectorial relation between DE vs. TE known as the Rasmussen  classification discussed in   that expresses the degree in which one industry output is used by other industries as an input. In this case, this grouping is based on the comparison of the median values of the sectorial DE and TE in a logarithmic scale. Table 5 shows the meaning of each region.
Table 2. Baseline scenario.
A large discussion of Rasmussen method is developed in  . It corresponds to a Classical Multiplier Method   . Although there are new developments in the methods developed to analyze interlinkages among industrial sectors, this method is very useful in identifying total and distribution effects, particularly in the analysis of the economic impacts of GHG mitigation    .
The sectors located in region I of Figure 1 are the construction and automotive sectors. These sectors use inputs of other productive processes, that is to say their consumption is influenced by the demand of other sectors. Consequently, mitigation policies that could affect the magnitude of their production might generate problems in their economic activity. In addition, changes in automotive industries’ (automotive production) final demand have a small influence on total
Table 3. Total effect (TE) among all sectors of the economy (%).
emissions, but the changes in the final demand of other sectors have large impacts on emissions, demonstrating the important influence of this sector on eco- nomic activity. A reduction in its final demand would have large impacts on economy and small impacts on emissions.
In Region II we can find the following sectors: road transport, electric power generation, brewages, chemistry, agriculture, iron and steel, commerce, oil and gas extraction, air transportation, and other sectors. Changes in final demand of these specific sectors have a large influence on total emissions, and changes in final demand of other sectors also have large impacts on emissions of these specific sectors. A demand reduction in these sectors will have a large influence on emissions, but also might have a large influence on economic activity.
Table 4. Distributive effect (DE) among all sectors of the economy (%).
Figure 1. Distributive effects vs. total effects.
Table 5. Regions in Figure 1.
The sugar industry and the cement industry are the sectors in Region III. Changes in final demand of these specific sectors have a large influence on total emissions, but changes in final demand of other sectors have small impacts on emissions of these specific sectors. In Region IV are less relevant sectors in terms of final demand and emissions. A reduction in CO2-eq emissions of these sectors will not have an important impact on overall emissions, because the share in the distribution of emissions is low.
Another important observation from Figure 1 is how construction and cement (in region III) are linked. It is possible to connect a line with both sectors that crosses the mean values (the center of the graphic). TE of the construction sector that affects the cement sector is the same amount as the DE of the cement sector received from the construction sector. Hence, if the final demand of sector 21 decreases, the emissions from sector 14 will also decrease. This relation also means that if the cement for construction is substituted with other materials, emissions from sector 14 will decrease.
5. Concluding Remarks
In this paper, an input-output methodology is developed to analyze energy-re- lated GHG emissions of the Mexican economy. The paper also analyzes total and distributive effects that final demand has on emissions levels. It also identifies Own Sector Effects (OE) that result from the changes in the final demand of each Own Sector (the diagonal elements of matrix), and Structure Effects (SE) that result from the changes in other sectors of the economy.
According to IPCC’s fifth assessment report  , the main mitigation strategies for the industrial sector are 1) reduction of emission intensity expressed as the ratio of GHG emissions to energy use; 2) reduction of energy intensity, measured as unit energy consumption in physical units (or in this case monetary units); 3) increase in material efficiency, which is the amount of material required to produce one product; and 4) reduction of product service intensity, which is the level of service provided by a product.
These strategies can be applied to sectors that appear in Region II and III to obtain the largest reduction in GHG emissions. Strategies 3) and 4) are related to a reduction in material or product production and will have an important effect on the economy, particularly in those sectors that appear in Region II. The alignment of strategies to fulfill the goals of the NDC requires additional analysis. Additional work is necessary to evaluate policies. The results presented in this paper are a useful tool for a GEM for the Mexican economy.
Table A1. CO2-eq emission impact matrix for the year 2012(10-6%).
 Rogelj, J., den Elzen, M., Höhne, N., Fransen, T., Fekete, H., Winkler, H., Schaeffer, R.,Sha, F., Riahi, K. and Meinshausen, M. (2016) Paris Agreement Climate Proposals Need a Boost to Keep Warming Well Below 2℃. Nature, 534, 631-639.
 Markandya, A., Halsnaes, K., Lanza, A., Matsuoka, Y., Maya S., Pan J., Shogren, J., Seroa de Motta, R. and Zhang, T. (2001) Mitigation. In: Metz, B., Davidson, O., Swart, R. and Pan, J., Eds., Contribution of Working Group III to the Third Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, UK.
 Alcántara, V. and Padilla, E. (2009) Input-Output Subsystems and Pollution: An Application to the Service Sector and CO2 Emissions in Spain. Ecological Economics, 68, 905-914.
 Tarancón, M.A., del Río, P. and Callejas Albiñana, F. (2019) Assessing the Influence of Manufacturing Sectors on Electricity Demand. A Cross-Country Input-Output Approach. Energy Policy, 38, 1900-1908.
 Tarancón, M.A. and Del Río, P. (2012) Assessing Energy-Related CO2 Emissions with Sensitivity Analysis and Input-Output Techniques. Energy, 37, 161-170.
 Alcántara, V., Tarancón, M.A. and del Río, P. (2013) Assessing the Technological Responsibility of Productive Structures in Electricity Consumption. Energy Economics, 40, 457-467.
 Chen, G.Q., Guo, S., Shao, L., Li, J.S. and Chen, Z.M. (2013) Three-Scale Input-Output Modeling for Urban Economy: Carbon Emission by Beijing 2007. Communications in Nonlinear Science and Numerical Simulation, 18, 2493-2506.
 Cui, L.B., Peng, P. and Zhu, L. (2015) Embodied Energy, Export Policy Adjustment and China’s Sustainable Development: A Multi-Regional Input-Output Analysis. Energy, 82, 457-467.
 Chen, G., Hadjikakou, M. and Wiedmann, T. (2016) Urban Carbon Transformations: Unravelling Spatial and Inter-Sectoral Linkages for Key City Industries Based on Multi-Region Input-Output Analysis. Journal of Cleaner Production.
 Li, Z., Pan, L., Fu, F., Liu, P., Ma, L. and Amorelli, A. (2014) China’s Regional Disparities in Energy Consumption: An Input-Output Analysis. Energy, 78, 426-438.
 Wiebe, K.S., Bruckner, M., Giljum, S. and Lutz, C. (2012) Calculating Energy-Related CO2 Emissions Embodied in International Trade Using a Global Input-Output Model. Economic Systems Research, 24,113-139.
 Su, B. and Ang, B.W. (2013) Input-Output Analysis of CO2 Emissions Embodied in Trade: Competitive versus Non-Competitive Imports. Energy Policy, 56, 83-87.
 Cortés-Borda, D., Guillén-Gosálbez, G. and Jiménez, L. (2015) Assessment of Nuclear Energy Embodied in International Trade Following a World Multi-Regional Input-Output Approach. Energy, 91, 91-101.
 Bauer, N., Mouratiadou, I., Luderer, G., Baumstark, L., Brecha, R.J., Edenhofer, O., et al. (2013) Global Fossil Energy Markets and Climate Change Mitigation—An Analysis with REMIND. Climate Change, 136, 69-82.
 Sue Wing, I.N.J. and Timilsina, G.R. (2016) Technology Strategies for Low-Carbon Economic Growth: A General Equilibrium Assessment. Report No. 7742, The World Bank, Washington DC.
 Wiedmann, T., Lenzen, M., Turner, K. and Barrett, J. (2007) Examining the Global Environmental Impact of Regional Consumption Activities—Part 2: Review of Input-Output Models for the Assessment of Environmental Impacts Embodied in Trade. Ecological Economics, 61, 15-26.
 Alcántara, V. and Padilla E. (2003) “Key” Sectors in Final Energy Consumption: An Input-Output Application to the Spanish Case. Energy Policy, 31, 1673-1678.
 Wang, Y., Wang, W., Mao, G., Cai, H., Zuo, J., Wang, L. and Zhao, P. (2013) Industrial CO2 Emissions in China Based on the Hypothetical Extraction Method: Linkage Analysis. Energy Policy, 62, 1238-1244.
 Chang, N. and Lahr, M.L. (2016) Changes in China’s Production-Source CO2 Emissions: Insights from Structural Decomposition Analysis and Linkage Analysis. Economic Systems Research, 28, 224-242.
 Lamonica, G.R. and Mattioli, E. (2015) Research Note: The Impact of the Tourism Industry on the World’s Largest Economies—An Input-Output Analysis. Tourism Economics, 21, 419-426.
 Fischedick, M., Roy, J., Abdel-Aziz, A., Acquaye, A., Allwood, J.M., Ceron, J.-P., Geng, Y., Kheshgi, H., Lanza, A., Perczyk, D., Price, L., Santalla, E., Sheinbaum, C. and Tanaka, K. (2014) Industry. In: Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler, A., Baum, I., Brunner, S., Eickemeier, P., Kriemann, B., Savolainen, J., Schlömer, S., von Stechow, C., Zwickel, T. and Minx, J.C., Eds., Climate Change 2014: Mitigation of Climate Change, Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge and New York, 739-810.