Report No: ACS3784 . United Mexican States Reducing Fuel Subsidies: Public Policy Options . May 2013 . LCSEN LATIN AMERICA AND CARIBBEAN . . . . Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. . Copyright Statement: . The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. CURRENCY Currency Unit = $/. Peso US$1 = S/. 12.172900 FISCAL YEAR January 1 to December 31 Vice President: Hasan Tuluy Country Director: Gloria Grandolini Sector Director: Ede Ijjasz-Vasquez Sector Manager: Karin Kemper Sector Leader: Alexandra Ortiz Task Manager: Ernesto Sánchez-Triana 4 Table of Contents Acronyms ................................................................................................................................................... 7 Acknowledgements .................................................................................................................................... 8 Executive Summary 6 Introduction .............................................................................................................................................. 12 1. The Dynamic Computable General Equilibrium Model .................................................................. 14 2. Scenarios and Results ....................................................................................................................... 18 3. Environmental Impacts of Reforms ................................................................................................. 39 4. Policy Implications........................................................................................................................... 40 5. References ........................................................................................................................................ 40 Appendix .................................................................................................................................................. 44 A. The model and recent applications B. Building in the Stylized Facts: Business as Usual vs. Benchmark C. International comparison of subsidies across selected countries D. Trade under Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual........................... E. Partial Neutrality F. Non-Neutral Elimination of Energy Subsidies and Expansion of VAT to Food and Medicines G. Universal Social Insurance Proposal H. Trade under Dual Policy Implementation vs. Business as Usual I. International Comparisons of CO2 Emissions per GDP 5 TABLES: Table 1: Household Categories Based on Income 11 Table 2: Producing Sectors and Consumption Goods 12 Table 3: Subsidies to Electricity (% of price), 2005–10 18 Table 4: Subsidies to Gasoline (% of price), in 2010 19 Table 5: Change in Aggregate Results, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 20 Table 6: Change in Production, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 24 Table 7: Change in Consumption, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 26 Table 8: Change in Aggregate Results, Neutral Subsidy Phaseout vs. Business as Usual 25 Table 9: Change in Production, Neutral Subsidy Phaseout vs. Business as Usual 26 Table 10: Change in Consumption, Neutral Subsidy Phaseout vs. Business as Usual 229 Table 11: Change in Aggregate Results, Expanded Healthcare vs. Business as Usual 30 Table 12: Change in Production, Expanded Healthcare vs. Business as Usual 31 Table 13: Change in Consumption, Expanded Healthcare vs. Business as Usual 32 Table 14: Change in Aggregate Results, Joint Policies vs. Business as Usual 34 Table 15: Change in Consumption, Joint Policies vs. Business as Usual 34 Table 16: Change in Production, Joint Policies vs. Business as Usual 35 Table 17: Burden of Health Impacts of Urban Air Pollution 37 FIGURES: Figure 1: Business-as-Usual Case 17 Figure 2: Gradual Subsidy Removal: Non-Revenue Neutral 22 Figure 3: Eliminating Energy Subsidies and Financing Expanded Healthcare 36 6 Acronyms CES Constant Elasticity Of Substitution CGE Dynamic Computable General Equilibrium CO Carbon Monoxide CO2 Carbon Dioxide CO2e Carbon Dioxide Equivalent DAC High-Consumption Tariff GDP Gross Domestic Product IEPS Impuesto Especial sobre Produccion y Servicios INEGI Instituto Nacional de Estadística y Geografía LPG Liquefied Petroleum Gas MPSGE Mathematical Programming System for General Equilibrium MtCO2e CO2 equivalent PECC Special Program on Climate Change (Programa Especial de Cambio Climático) SAM Social Transactions Matrix SHCP Ministry of the Treasury (Secretaría de Hacienda y Crédito Público) SO2 Sulfur Dioxide VAT Value-Added Tax 7 Acknowledgements The report “Mexico: Reducing Subsidies to Energy Use and Analyzing Alternative Compensation Mechanismsâ€? was prepared by María Eugenia Ibarrarán (Consultant), Roy Boyd (Consultant) and Alejandra Elizondo (Consultant). The task team included Santiago Enriquez, Francisco Giner de los Rios, Juan C. Belausteguigoitia, Santiago V. Sandoval, Geise B. Santos, and Ernesto Sánchez-Triana (Task Team Leader). The team would like to thank the Ministry of the Treasury (Secretaría de Hacienda y Crédito Público), especially Emilio Pineda and Luis Octavio Alvarado (Subsecretaría de Ingresos, Unidad de Política de Ingresos), for their data on subsidies; Arturo Antón and Fausto Hernández-Trillo for the discussion of social policies; Xian Fo for his support as a research assistant throughout the preparation of this study, and Santiago Levy for his helpful comments and discussion. The team also wants to thank comments by the Ministry of Health, particularly Vice-minister Igor Rosete for his insightful remarks and by the peer reviewers Carlos Hurtado, Todd Johnson and Muthukumara Mani. The study was carried out under the overall guidance of Gloria Grandolini (Country Director), Ede Jorge Ijjasz-Vasquez (Sector Director), Karin Kemper (Sector Manager), and Alexandra Ortiz (Sector Leader). Financial support by the Spanish Fund for Latin America (SFLAC) is gratefully acknowledged. 8 Executive Summary This paper analyzes the economic, distributional, and environmental impact that energy subsidy reductions and alternative compensating mechanisms might have in Mexico1. To achieve that goal, we use a computable general equilibrium model of the Mexican economy (for a detailed description, see Ibarrarán and Boyd 2006). We build the energy subsidies (to gasoline, diesel, and electricity) into the benchmark and then do an array of simulations to see the effects of removing such subsidies. We report results for 2012, which is the initial year; 2018, which would be the end of the next administration; and 2024 and 2030, which represent the medium and long term, respectively. When doing the simulations, we look at possible compensation mechanisms and analyze the impact on the income groups that may be affected by the reduction of energy subsidies. Energy subsidies are a politically sensitive topic in Mexico. They include subsidies to electricity, gasoline, diesel, and liquefied petroleum gas (LPG). Between 2005 and 2009, subsidies were, on average, equal to Mex$200.4 billion per year (SHCP 2010). Subsidies to electricity alone represent roughly 1 percent of GDP. Subsidies to gasoline are equivalent to 25 percent of revenues collected from the value-added tax (VAT). The amount of money used toward these subsidies has increasingly been questioned because of their fiscal, environmental, and distributional implications. In addition, energy subsidies have not given the correct incentives toward energy efficiency, and have backfired in terms of local and global emissions, contributing to health problems and climate change. However, because eliminating these subsidies is a politically charged issue, concrete alternatives must be proposed to use those resources toward other policies that may be perceived as welfare enhancing. One possible compensation mechanism in addition to many others is to reallocate the funds to finance a universal healthcare program. Several authors have proposed alternative packages (Antón, Hernández, and Levy 2012; Chávez Presa, Hernández Trillo, and López-Calva 2012, among others). Toward that end, we examine the implications of financing a expanded healthcare system with a cost of approximately Mex$560 billion a year. Depending on the specifics of the proposal, this might entail some combination of healthcare coverage, life insurance, and/or pension systems. The more ample the package, the higher its cost, so ultimately it will be up to the social security reform experts to put together the best combination of policies they can finance. The exercise presented here does not advocate for any particular package; rather, it presents to what extent a package of certain cost can be financed with the savings from energy subsidies. At the same time, Mexico is a growing emitter of greenhouse gases. Even though it is ranked 13th worldwide and its emissions are only about 1.6 percent of total global emissions (CICC 2009), Mexico contributes approximately 25 percent of the emissions of Latin America and the Caribbean. Since Mexico is an upper-middle income developing country that is interested in becoming a significant player in climate change policy worldwide, this makes emission abatement policy relevant. In 2006, 28 percent of greenhouse gas emissions came from energy generation and 33 percent from energy use. Energy subsidies play a large role in this. Of total emissions from energy use, 62 percent are emitted by the transport sector and 10 percent by the residential sector (CICC 2009). Both of these sectors are 1 Energy subsidies are measured as the sales price with respect to production costs for each energy source. In the case of electricity, domestic production costs are used to estimate the subsidies; for all other fuels, international prices are taken into account to estimate the reference cost. No externalities are taken into account. The numbers used for the subsidies were provided by SHCP. 9 heavily subsidized. The Mexican government has put forth a plan to mitigate emissions, the Special Program on Climate Change (Programa Especial de Cambio Climático, PECC), running from 2009 to 2012. Overall, it estimates that 129.03 million tons of CO2 equivalent (MtCO2e) would be eliminated during that four-year period, 50.65 in 2012 alone. Subsidy elimination was not considered within the PECC. Additionally to climate change mitigation commitments, Mexico faces severe problems related to air pollution and groundwater exhaustion. All of these problems are exacerbated by energy subsidies that allow for lower cost fuels that cause pollution and promote water extraction through subsidized electricity tariffs to the agricultural sector. Thus, eliminating energy subsidies may help address many of these problems simultaneously. The use of a computable general equilibrium model of the Mexican economy allows us to analyze the economic, distributional, and environmental impact of energy subsidy reductions. We first build the benchmark scenario and then we include some stylized facts of the economy (such as slower growth in the production of oil, unemployment, and formal and informal work in various sectors); this is the Business as Usual Scenario. We then simulate the elimination of subsidies to energy, and the Government retains those saved resources. There are negative short term effects both in GDP and investment, positive effects in aggregate GDP of the order of 1.5%, mainly explained by a significant increase in investment by the Government, and growth in the production and consumption of the various sectors and assets in the long term. On the other hand, this increase in investment leads to a progressive and positive effect on well-being of the various income groups, only affecting negatively the highest income groups, i.e. which are heavy consumers of energy. The short run negative effects of subsidy elimination may be partially compensated using different schemes to transfer those saved resources to the population, sometimes to all, others to low-income groups, and sometimes only part of the resources saved from the subsidies. This in itself avoids a stagnation of economic activity and keeps up both expenditure and demand. We also provide results for these simulations and discuss some of the sensitivity analysis we carry out. In any case, over these runs, the energy-intensive sectors are significantly affected because now they do not receive the subsidy. Effects are quite similar under a gradual and a sudden removal of subsidies, given the same compensation mechanisms. We then simulate the effect of expanding healthcare. Healthcare is not readily available to all workers in Mexico, and those that are not covered by any of the existing health systems face catastrophic expenditures and expensive private services if they happen to need medical attention. We run an expanded healthcare scenario where we first look into expanding the VAT to food and medicines since they are currently not covered by that tax2 and from a fiscal standpoint this is a first step toward any tax reform. Additionally, the purpose of building this into the model is to partially finance the expanded healthcare coverage. The main results in this case are a significant increase in GDP, in the order of 2.8% by the end of 2030, growth in investment and capital accumulation, significant and progressive gains in welfare, and an increase in production and consumption across the board. The 2 Even though this is somehow not central to the central argument of this paper, VAT on food and medicine is bound to be a central component of any fiscal reform to close loop holes, for its progressivity and revenue raising capacity, and for the clarity of the tax system, so we include it as a building block to the expanded healthcare scenario. This increased coverage of VAT has been debated in Mexico under recent years, and is thus politically relevant. 10 universal healthcare program promotes formality in the Mexican economy and leads to higher productivity, with a more equitable income distribution. We then provide some sensitivity analysis for different rates of substitution among formal and informal workers across sectors. Finally, when energy subsidies are eliminated and the expanded healthcare programs are combined, the results are highly desirable. For example, growth in GDP with respect to the Business as Usual case is in the order or 3%. This leads to the conclusion that these reforms help reduce distortions to the economy; increase its productivity and have positive effects on growth, investment, and capital accumulation; promote a more equitable income distribution; and reduce emissions that have both global and local effects. Eliminating subsidies to energy achieves several goals. It cuts back on the use of a valuable non- renewable resource and the extraction costs of which are rising, and it promotes energy savings and efficiency. This, in turn, reduces emissions that have both global and local effects, contributing to improved air quality and reduced health costs. In addition, this contributes to mitigating climate change, since subsidy cutbacks alone get 80 percent of Mexico’s yearly abatement goals for CO2 emissions, as set by the 2008–2012 Special Program on Climate Change. Lower CO2 emissions also imply lower local emissions, leading to improved environmental quality across the board. This is relevant since air pollution imposes a yearly cost of 1.5 percent of GDP on Mexico, mainly measured through urban impacts. Finally, since a significant part of energy subsidies goes to pumping water for agriculture, eliminating them may also help reduce groundwater extraction and therefore the replenishment of such sources. In summary, this exercise shows that the proposed reallocations of resources used to subsidize energy produce higher welfare (especially for low-income groups), promote overall economic growth, and improve environmental conditions and health. 11 Introduction 1. This paper analyzes the economic, distributional, and environmental impact that energy subsidy reductions and alternative compensating mechanisms might have in Mexico. To achieve that goal, we use a computable general equilibrium model of the Mexican economy (for a detailed description, see Ibarrarán and Boyd 2006). We make several important changes to the original model to build the energy subsidies (to gasoline, diesel, electricity and LPG) into the benchmark and then do an array of simulations to see the effects of removing such subsidies. We report results for 2012, which is the initial year; 2018, which would be the end of the next administration; and 2024 and 2030, which represent the medium and long term, respectively. 2. There can be basically an infinite array of alternative ways to make changes to the energy pricing policy. When doing the simulations, we first look at the elimination of energy subsidies, and then at different scenarios and several possible compensation mechanisms and analyze the impact on the income groups that may be affected by the reduction of energy subsidies. First we look at different types and magnitudes of lump-sum transfers, where all the saved subsidies are transferred to all of the population on a per capita basis, and then we consider variations on this, like only transferring half of the resources saved to the 50% lower income groups. Another option to use the saved funds is to finance an expanded healthcare program. Several authors have proposed alternative packages (Antón, Hernández, and Levy 2012; Chávez Presa, Hernández Trillo, and López-Calva 2012, among others). Toward that end, we examine the dynamic implications of financing a healthcare system with a cost of approximately Mex$560 billion a year3. This helps illustrate how those resources could be spent. Other proposals that entail different aspects and/or coverage could well be simulated. Depending on the specifics of the proposal, this might entail some combination of healthcare coverage, life insurance, and/or pension systems. The more ample the package, the higher its cost, so ultimately it will be up to the social security reform experts to put together the best combination of policies they can finance. The exercise presented here does not advocate for any particular package; rather, it presents to what extent a package of certain cost can be financed with the savings from energy subsidies. 3. On the other hand, Mexico is a growing emitter of greenhouse gases. Even though it is ranked 13th worldwide, it contributes approximately 25 percent of the emissions of Latin America and the Caribbean, although its emissions are only about 1.6 percent of total global emissions (CICC 2009). This makes emission abatement policy relevant since Mexico is a high-income developing country that is interested in becoming a significant player in climate change policy worldwide. In 2006, 28 percent of greenhouse gas emissions came from energy generation and 33 percent from energy use. Energy subsidies play a large role in this. Of total emissions from energy use, 62 percent are emitted by the transport sector and 10 percent by the residential sector (CICC 2009). Both of these sectors are heavily subsidized. The Mexican government has put forth a plan to mitigate emissions, the Special Program on Climate Change (Programa Especial de Cambio Climático, PECC), running from 2009 to 2012. Overall, it estimates that 129.03 million tons of CO2 equivalent (MtCO2e) would be eliminated during that four-year period, 50.65 MtCO2e alone in 2012. Subsidy elimination was not considered within the PECC. Results of the model presented here show that roughly 80 percent of CO2 emissions could be 3 As an illustration, we assume that the current cost in health per worker is $10,118 pesos. Given a workforce of 39.03 million workers, the cost of health insurance would be Mex$394.9 billion. To this amount, Mex$136.3 billion are added to cover retirement pensions and Mex$28.1 billion for life and disability insurance. This gives a total cost of Mex$559.3 billion according to Anton et al (2012). 12 abated only by the elimination of energy subsidies, significantly contributing to the mitigation goals stated under PECC. Other environmental goals can also be achieved such as pollution reduction and thus health costs and reducing groundwater overexploitation. 4. This paper is organized as follows. Section 1 describes the model, section 2 describes the scenarios and results, section 3 discusses the environmental impacts of these policies, and section 4 analyzes the findings from a public policy perspective. The Appendix contains supporting information helpful for the analysis and the discussion of the results. 13 1. The Dynamic Computable General Equilibrium Model 5. We turn to the issue of modeling the general equilibrium effects of eliminating energy taxes across the board, basically to gasoline, diesel, and electricity. These policies that affect large and interrelated sectors are not conducive to an analysis within a simplified framework, since they can have important repercussions throughout the economy. Such problems are appropriately dealt with using general equilibrium analysis. In this type of framework all the sectors in the economy are seen as one linked system in which a change in any part affects prices and output economywide. Mathematically, an interlinked economy can be described by a large system of simultaneous equations. More precisely in an economy with N markets, we require N-1 equations to solve for all of the prices and outputs in the system. 6. In this paper we look at a national model that has 12 producing sectors.4 The primary sector is disaggregated into agriculture, livestock, fisheries, and forestry (see Table 1). This was done so that we can now explicitly deal with and quantify the interaction of sector-specific policies with other sectors when policies are initiated. It is particularly important to do this given that the simulations we run affect across-the-board energy subsidies and several taxes to finance expanded healthcare, for example. The model has four household (income) categories (listed in table 1) and nine consumption sectors (in Table 2). There is also a foreign sector and a government in this model. This model uses the latest information from the input-output matrix produced by the Instituto Nacional de Estadística y Geografía (INEGI), which uses 2003 as the base year, and other data from both national and international sources. Table 1: Household Categories Based on Income Category Income Group Agent 1 Bottom 2 deciles: 1–2 Agent 2 Deciles 3–5 Agent 3 Deciles 6–8 Agent 4 Top 2 deciles: 9–10 Source: Authors. 7. The economic variables determined by the model are investment, capital accumulation, production by each sector, household consumption by sector, imports and exports, relative prices, wages and interest rates, government budget expenditures and revenues, and total wage income. The level of depreciation and the initial return to capital are taken as exogenous, as is the rate of labor force growth. For a detailed explanation of the model and its recent application, see Appendix A. 4 Data restrictions prevent us from constructing a regional model. Furthermore, a regional model is impractical given that the capital in any given region is owned by individuals and corporations throughout the country. Lack of regionalization is not a major drawback, however. 14 Table 2: Producing Sectors and Consumption Goods Producing Sectors Consumption Goods Agriculture Food Livestock Household and other goods Fisheries Consumption services Forestry Energy (electricity and LPG) Manufacturing Autos Chemicals & plastics Gasoline Mining Public transport Oil and gas Housing Transport Water Electricity n.a. Services n.a. Refining n.a. Source: Authors. Note: n.a. = Not applicable. Production5 8. In each time period producers maximize profits in a competitive environment. Profit maximization, based on the described production technology, yields output supply and factor demands for each production sector and factor market in the model. Output and input prices are treated as variables. Taxes are also included, with producers facing tax-exclusive prices and consumers (and input-consuming firms) facing tax-inclusive prices. 9. As a word of caution, the goods produced in the model’s production sectors are not the same final goods consumed by consumers. Agricultural products, for example, must be combined with transportation services, manufacturing, and chemicals before they can be consumed by individuals as food. Hence, in our model we use a matrix to map from the vector of production goods to the vector of consumption goods. We do this through the use of nested functions to the production side of the economy and to the production of final consumption goods and services. This allows for different degrees of substitution for the inputs considered, particularly between labor, capital, energy, and non- energy inputs. Technologies are represented by production functions that exhibit constant elasticities of substitution. Technical progress is taken as exogenous to the model. Consumption and Income Distribution 10. On the demand side, the model reflects both the behavior of domestic consumers and foreigners (who can also invest through their savings) and that of the government. Domestic consumers are assigned to four groups (agents) according to income, and a demand equation is specified for each group, which has a different consumption bundle depending on its income. All four groups are 5 For a formal mathematical description of the model, see Ibarrarán and Boyd (2006, 114 –26). 15 endowed with labor. Since only the wealthy actually have (formal) savings in Mexico, we assume here (in accordance with the latest data from INEGI) that only the top two groups (agents 3 and 4) own capital.6 The gross income of each group rises by the rate of population growth plus the rate of technological change, which is taken as capital augmenting. These resources are rented out to firms in order to finance the purchase of domestic or foreign goods and services, to save, or to pay taxes to the government. The membership of each group is fixed, and although group income increases (or decreases) with GDP, individuals do not “migrate,â€? as such, from group to group.7 Government 11. The government agent is modeled with an expenditure function similar to the household expenditure functions (that is, based on a constant elasticity of substitution [CES] utility function). Revenues derived from all taxes and tariffs are spent according to an expenditure function. Within this expenditure function the government spends its revenues on goods and services from the various private production sectors discussed above. Consistent with the treatment of Ballard et al. (1985) and others, we posit an elasticity of substitution between inputs to the government’s utility function. This allows for price responsiveness in the provision of government-purchased goods. The government also spends its revenues on labor. Together, these arguments represent the government purchases and payment of government employees necessary for it to carry on its work. The government also separately redistributes income through exogenously set subsidies and transfer payments, and the government budget is assumed to be balanced. 12. Taxes in the model are expressed ad valorem and include personal income taxes, labor taxes, capital taxes, property taxes, revenue taxes (such as payments from oil and gas activities), value-added taxes, sales taxes, and import tariffs. The taxes on final goods such as gasoline differ from other consumer goods because of special taxes levied on them by the government, that is, the Impuesto Especial sobre Produccion y Servicios (excise tax, IEPS). By the same token, final goods such as electricity differ in treatment due to existing government subsidies. Income taxes are based on marginal tax rates. Subsidies to industries are essentially treated as negative taxes, and in these cases the government transfers funds back to a sector in proportion to that sector’s output. Thus, if these subsidies are abolished, the government has more revenue. Trade 13. International trade within the model is handled by means of a foreign agent. Output in each of the producing sectors is exported to the foreign agent in exchange for foreign-produced imports. Under this setup, the aggregate level of imports is set and grows at the steady state level, but the level of individual imports may change in response to changes in relative prices. Exports are exogenous, as well, and are assumed to follow a constant growth path. They are, however, responsive to changing prices, and can change as individual sectors are shocked. Transfer payments, however, are endogenous 6 Household savings here have a certain degree of endogeneity. The level of savings for each income group (that is, Agents 3 and 4) are set at the levels that actually occurred in the base year of the dataset (that is, 2004). After that time, however, they are allowed to vary in response to changes in the relative prices of consumption and savings. 7 Such migration, though a concept to explore, is computationally beyond the scope of this model. Furthermore, our chief concern with income distribution is how different income groups with varying consumption bundles and income streams are differentially impacted by the effects of policies. 16 and act to clear the model. The exchange rate is determined then by the interaction of capital made available for external uses, goods supplied for export, and the exogenous level of imports.8 Price- dependent import supply schedules are derived from elasticity estimates found in the literature. 9 In specifying the level of substitutability between goods, we rely on the Armington (1969) assumptions, which allow for imperfect substitutability between foreign and domestically produced goods. 14. In this model, we assume that Mexico has no market power in the world petroleum market. Hence, we treat the international price of oil as a given, and Mexican oil producers as price takers in the market. Consequently, when the Mexican government institutes investment policies to increase aggregate oil output, the domestic price drops as output increases and more is exported as the international price increases relative to the domestic price.10 Oil depletion, however, represents a curbing investment. Labor Growth and Capital Formation 15. Growth within our Dynamic Computable General Equilibrium (CGE) model is brought about by the changes over time in both the labor force and the capital stock. In keeping with the theoretical underpinning of the Ramsey model (1928), we take the changes in the population as exogenous and constant over the time period considered. In the absence of any perturbation, Ramsey predicts that the economy will grow at the labor supply growth rate in the steady state. 16. In the model, we assume that there is only one type of raw capital good, which goes into the various sectors. In addition, to add realism, we assume that the capital that goes into a sector works like putty and clay. More specifically, we assume that capital that is new can be readily combined with other inputs to produce outputs. Over time, this capital becomes locked into an older technology (that is, clay) and has a harder time combining with other inputs. In the growth literature, this is also known as “vintage capital.â€? This is plausible as illustrated by sectors such as electricity production, which has been subject to a great deal of technological change over the years. The capital growth rate is modeled in accordance with neoclassical capital theory assumptions. More specifically, the growth of capital is modeled as investment net of economic depreciation. Benefits of a computable general equilibrium model 17. Some sectors are crucial to the way an economy works. Energy is one of such sectors, so any changes to pricing policies such as subsidy reductions will have effects over the entire economy. The use of a computable general equilibrium model, that is, a framework where all the sectors in the economy are seen as one linked system in which a change in any part affects prices and output economy-wide, is highly recommended, as opposed to a simplified framework of just one sector (i.e. partial analysis) that can have important limitations since no interaction effects may be captured. Economy-wide effects are appropriately dealt with a computable general equilibrium model specially designed for Mexico that is able to capture all the interrelations across the different sectors and on the 8 As a side note, closure in our model is determined by the equality of domestic and foreign leakages and domestic and foreign injections. More formally, we have (S + M) = (I + X) where S is domestic savings, M is imports (current account), X is exports (current account), and I is the total amount of investment made available from foreign and domestic agents. 9 See, for example, Fernandez (1997); Romero (1994); Serra-Puche (1984); and and Wylie (1995). 10 The domestic and international prices of oil may differ due to quality and transportation costs. 17 different consumer groups. Additionally this is a dynamic model, so there is growth and capital accumulation along the period of analysis. This allows to a better understanding of how different variables adjust through time and reflects the way the economy works. It enables making complex simulations on different policies that may alter the long-term situation of the economy. Some applications of the same basic model used in this paper are described in Appendix A. Caveats 18. Even though these models are very powerful, the usefulness of the results depends basically on the quality of the information that is fed, such as data from the input-output matrix and technical parameters of substitution, price and income elasticities, and other national data. On the other hand, even though this information is accurate, there is uncertainty about the exact value of the parameters, and of certain assumptions such as long term growth and population dynamics, for example, among other key variables. Therefore, the results should be interpreted with caution. For example, the basic take-home points should be the sign, the trend of the different variables, and the relative magnitude of the resulting numbers more than the values themselves. These results then show if the policy has positive effects, if they grow in time and what sectors are affected relatively more. 19. On the other hand, this is a national level model that has no regional interpretation. To have a regional model we would need regional input-output data and social accounting matrices. This, however, is not a drawback in this particular exercise given that the energy sector, and labor for that matter, across all sectors. 2. Scenarios and Results 20. This section describes the various scenarios we ran for this project and their comparison. We start by building the subsidies into the benchmark and then adding selected stylized facts of the Mexican economy, such as oil depletion and unemployment. We then simulate elimination of energy subsidies together with different types of lump-sum transfers to the population. Since lump-sum transfers are difficult to achieve because of the practical complications of getting the money to the people, we looked for alternative public policies to finance so the energy subsidies elimination is revenue neutral. One option among many is an expanded healthcare policy to cover employees in the formal and informal sectors under equivalent healthcare programs. To achieve this, we model a partial removal (about two-thirds) of social security employee-employer contributions, and the expansion of the VAT to food, medicines, and medical services11. We run this scenario under different elasticities of substitution for labor across the formal and informal sectors. Finally, we combine the elimination of energy subsidies with the expanded healthcare program, supported by some tax changes 12. 11 These are currently not covered under VAT, but there has been an ongoing discussion in Mexico on the option of expanding such coverage to these categories within any future fiscal reform. 12 The reader may decide to only focus on scenarios 3 and 6 to see the effect of energy subsidy elimination alone and the joint implementation of both policies under study, i.e. elimination of energy subsidies and an expansion of the healthcare system financed by both saved energy subsidies and an extended VAT to food and medicines. For detailed explanations, review the appropriate scenarios. 18 Scenario 1. Benchmark 21. Initially, we want to make sure that our CGE model is balanced both in terms of the social transactions matrix (SAM) constructed and in terms of its dynamic capital and labor components. To do this we run the model without any policy changes or dynamic constraints whatsoever. If the model is properly constructed, this should result in an algebraically consistent replication of the 2004 economy growing at the pre-specified rate of economic growth of 3.2 percent per year and at the pre- specified level of capital depreciation and investment per year. Doing this, we find that the model is indeed specified correctly and that the Mexican economy is growing at the proper “steady state.â€? Results for this simulation are not shown (given their lack of intuition for policy analysis), but can be obtained from the authors upon request. Scenario 2. The Business-as-Usual Case 22. This balanced “steady stateâ€? run, though computationally important to our analysis, is unrealistic and of no use to the policy questions we are trying to answer. It assumes that the rate of growth is constant in all sectors including nonrenewable extraction of oil and natural gas. It also assumes no unemployment. 23. Thus, to begin our formal analysis, we assume that the level of oil extraction is initially approximately 2.6 million barrels per day and follows Secretaría de Energía de México (Ministry of Energy) estimates (SENER 2012), flattening out at approximately to 3.3 million barrels per day. Natural gas goes from 6.2 million cubic feet to 8.7 million cubic feet. Unemployment is included in the model at 3.3 percent (frictional) in the initial year, that is, 2004. The aggregate results of the comparison between business as usual and the benchmark are presented in Appendix A1. They are not listed below because they are not directly relevant to the policy analyses proposed in this exercise. However, we present the dynamics of the different sectors once all the stylized facts described above are incorporated. Figure 1 shows how investment and selected sectors grow relative to each other under this scenario. The driving force behind lower growth in some sectors is oil depletion. 19 Figure 1: Business-as-Usual Case 2.40 Percent increase in each sector relative to the 2004 2.20 Forestry 2.00 Mining 1.80 Transportatioin baseline year Electricity 1.60 Pet&NG 1.40 Services Investment 1.20 Manufacturing 1.00 Refining Agriculture 0.80 Chemicals Year Source: Authors. 24. Looking at our results for production and consumption, we find that there is a general decline in activities in all sectors. Growth is even more negative in production sectors such as electricity and refining, which are closely related to the burning of fossil fuels. Growth is also negative in consumer sectors such as energy and gasoline. 25. As for the foreign sector, oil and natural gas and refining sector imports grow to compensate for the domestic exhaustion of oil. The rest of the sectors are not significantly affected in terms of imports and exports. In all cases, the business-as-usual case is crucial since all of the following scenarios will be compared against it. The main results are presented in Appendix B. Scenario 3. From Business as Usual to Non-Neutral Gradual Removal of Energy Subsidies 26. In Scenario 3, we look at the effect on the Mexican economy of gradually removing energy subsidies. The amount of subsidies was provided by the Secretaría de Hacienda y Crédito Público (Ministry of Finance). The subsidies (expressed as a share of price) to electricity and gasoline are shown in tables 3 and 4, respectively. The electricity consumption of the residential sector is divided into a highly subsidized category (T1 and TF1 may have subsidies above 200 percent), and a high- consumption tariff (DAC) that is much less subsidized or even taxed, depending on the year analyzed. However, the DAC is less than 5 percent of residential power consumption13. 13 Subsidies to electricity were provided by SHCP and they are subsidy/price. In the case of electricity, the size of a subsidy is not easy to determine as it implies dealing basically with a non-tradable good. The subsidy estimate is based on a 20 Table 3: Subsidies to Electricity (% of price), 2005–10 Electricity Subsidy (% of price) Sector 2005 2006 2007 2008 2009 2010 Residential 157.8 145.3 132.6 177.8 179.4 149.4 T1–T1F 190.2 182.9 183.7 245.2 214.6 182.7 DAC 2.6 -1.0 2.0 9.6 8.5 -16.7 Commercial 14.1 6.7 -2.9 13.6 19.0 -4.5 Services 30.4 28.8 25.0 44.9 40.8 14.4 Street lighting 26.4 25.8 24.2 42.2 25.4 3.1 Water pumping 40.4 36.6 38.1 66.6 48.9 41.5 Agriculture 251.1 227.0 213.5 258.4 244.8 242.8 a 9–9M 163.5 96.5 65.4 67.8 60.0 61.4 Incentive rate 282.1 297.3 290.4 357.2 291.7 314.2 Industrial 14.9 9.9 8.7 8.7 31.6 4.2 Medium voltage 16.2 14.3 15.1 13.6 35.1 3.1 High voltage 11.0 1.1 1.8 -2.0 23.5 1.5 27. Gasoline prices are fixed by the government but with no transparent rule. This leads to some years in which it is highly subsidized and other periods when it is taxed14. We use 2010 as the reference year to determine subsidies that again are expressed as a share of the price (to see some international comparisons see Appendix C). These are the subsidies that we eliminate in the next simulations15. calculation of the longer-term marginal cost of production with an administratively determined rate of return on assets. Inefficiencies in the production, transmission and distribution are part of the subsidy estimate and the latter is largely implicit as the government usually does not transfer a subsidy to the electricity company but largely foregoes a return on the assets it has historically invested. However, final price increases to consumer would have the effects described in this section. 14 For fuels, subsidies in all of the simulations are estimated by SHCP and they reflect the subsidy as a share of domestic price. Given that fuel prices in Mexico are fixed by the government, volatility in international prices (that are taken into account when estimating costs) some years lead to negative subsidies and some to positive subsidies. In 2010 we observed a positive subsidy and that is the level of subsidies we work on eliminating in this document, 15 Even though there may be price volatility in gasoline and oil prices, we take the subsidies in 2010 as a possible example. Other options could have been used such as having the average subsidy over some time period, for example. A fact is that price volatility affects the size of subsidies. In any case, lower subsidies will most likely have an impact on energy efficiency as is shown in the paper, when particularly electricity and gasoline consumption falls as prices rise. 21 Table 4: Subsidies to Gasoline (% of price), in 2010 Gasoline Subsidy, 2010 Sector (% of price) Agricultural sector diesel 13.55 Marine diesel 10.68 Auto vehicles diesel 13.55 Auto vehicles gasoline 18.69 Agricultural sector gasoline 18.76 Source: Secretaría de Hacienda y Crédito Público, Subsecretaria de Ingresos – Unidad de Política de Ingresos. 28. Thus, in this simulation, we calculate the impact of removing the existing gasoline, diesel, and electricity subsidies on consumers, public transportation, agriculture, fisheries, public lighting, and various manufacturing industries. In our simulation, the subsidy removal begins in 2012 and is completed in 2018 to avoid any sudden shock to the economy. In addition, to concentrate solely on the substitution effects of the subsidy removal, the policy is done in a non-revenue-neutral manner, with all funds from subsidy removal being collected by the government and then invested according to its own welfare goals. To give us a frame of reference, we run the model as described with subsidy removal taking place along with fossil fuel depletion, and compare our results with those of the business-as-usual case described in Scenario 2, above. 29. Looking first at the aggregate numbers in Table 5, we see that removing energy subsidies increases the aggregate level of welfare by 0.84 percent 16. Aggregate GDP declines slightly initially (that is, in 2012) due to adjustment processes, but rebounds quickly and increases by 0.34 percent in 2018 and 1.54 percent by the final period of the analysis. Aggregate investment increases significantly in the latter part of the analysis, and the final level of the capital stock goes up by 7.13 percent as a result of new investment. The aggregate level of government welfare goes up by 3.5 percent and the welfare levels of the poorest agents go up by 1.10 percent and 1.01 percent, respectively, in spite of rising expenses as the gross domestic product (GDP) rises. Agent 3 sees a smaller increase of only 0.43 percent, and Agent 4 sees a slight decline of 0.18 percent. Thus, economic growth declines only in the very earliest part of the analysis and then speeds up in the later periods. A total removal of energy subsidies, it would seem at first glance, is beneficial to the level of economic welfare in Mexico over the 2004 to 2030 period studied. Welfare gains are larger for lower income agents because they consume proportionally less energy than the rich and therefore they lose less of their purchasing power once subsidies to energy are removed. 16 The results of each variable under each scenario are compared with the value obtained in the same year but under the business-as-usual case, for example, GDP for 2012 under the scenario of elimination of subsidies is compared with the value of GDP in 2012 under business-as-usual scenario. The variations that are reported in the text correspond to the percentage changes between these two scenarios. 22 Table 5: Change in Aggregate Results, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP -0.9359 0.3358 0.6884 1.5397 Investment -2.8067 -0.0347 3.3236 16.1049 Government17 6.2460 6.4166 5.8566 5.0945 Capital Stock — — — 7.1322 Aggregate welfare (∑Agent 1–4) — — — 0.2920 Agent 1 — — — 1.0967 Agent 2 — — — 1.0063 Agent 3 — — — 0.4340 Agent 4 — — — -0.1820 Government welfare — — — 3.5042 Aggregate welfare — — — 0.8403 Source: Authors. Note: — = results not reported. 30. Several other things are worth noting here. First, the total level of government income plus consumer welfare rises because a subsidy removal increases net welfare by eliminating the welfare loss. Because the elasticities of demand for gasoline and energy goods are relatively inelastic (that is, ranging from -0.28 for gasoline to -0.32 for residential electricity18), the revenue gains by the government are fairly substantial even though the aggregate welfare gains are quite modest. This is consistent with partial equilibrium theory, since taxing (subsidizing) a good with an inelastic demand generates (expends) a large amount of revenue while leading to a small deadweight welfare loss (see, for example, Atkinson and Stiglitz [1980] for an extensive review of this theory19). 31. Turning next to the individual sectors, we find that, with the exception of the initial period (when capital and labor are adjusting to the subsidy removal), some or most production sectors see increases throughout the analysis. As would be expected, declines occur in the fossil fuel and electricity sectors that are most negatively impacted by the subsidy removal. Indeed, our model projects that, in 2030, electricity production will decline by approximately 25.3 percent compared to the business-as-usual case. Even though this number is high, it is explained by the extent of the subsidy removed. Significant losses are also experienced by the petroleum, natural gas, chemical (which includes petrochemicals), and refinery sectors. Increases are generally seen in all other sectors. Because of the large increases in investment, manufacturing (the largest supplier of investment goods) 17 Government refers to the total expenditure that, under a balanced budget, we assume here it is equal to total income from tax revenue and sales of publically provided goods and services. Since the idea here is to see how this concept changes when different policies are simulated, it out of the scope of this paper to include how the overall deficit will behave once policies are enacted in terms of its long terms sustainability. What we want to show here is how this balance in government revenues (or expenditure) changes under different policies. 18 Our gasoline elasticity is taken from Reyes Martínez, Escalante, and Matas (2010), and our residential electricity demand is taken from Bernstein and Griffin (2006) (this study is for the United States). All estimates of elasticities for energy demand, however, are inelastic in nature. Commercial demand for electricity was assumed to be about 1 (which is consistent with estimates by Bernstein and Griffin [2006] and Berndt and Samaniego [1983] for Mexico). 19 While, strictly speaking, welfare generation in CGE models is measured by equivalent variation (see Ballard et al. 1985), the partial equilibrium results follow their general equilibrium counterparts here fairly closely. 23 goes up by almost 14 percent by 2030. Fairly substantial increases are also experienced in the massive services sector for most of the analysis (Table 6). Here is important to note that even though sectoral growth lags with respect to aggregate GDP growth, services and manufacturing have positive growth even in 2018. This pulls aggregate growth even though individual smaller sectors linger behind. Table 6: Change in Production, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Sector (%) (%) (%) (%) Agriculture -2.1029 -1.4120 1.3247 8.9637 Livestock -1.6484 -0.4198 2.5197 10.1736 Forestry -0.8658 0.3597 3.2738 10.6729 Fisheries -2.0134 -2.2727 -0.9524 2.3622 Oil -5.5083 -8.4084 -6.5328 0.2007 Natural gas -5.4565 -8.3945 -6.4797 0.2523 Mining -3.6179 -2.3143 1.9057 12.7644 Refining -9.1142 -14.9005 -12.2972 -3.9018 Transport -1.2193 -1.0664 0.4664 4.1557 Electricity -24.5343 -35.6061 -32.7927 -25.3132 Chemicals & plastics -5.5546 -5.7596 -0.5537 14.0324 Services -0.2402 0.2896 1.7343 5.2546 Manufacturing -1.8774 0.0973 4.0762 13.9500 Source: Authors. 32. Figure 2 shows how particular sectors and investment behave once energy subsidies are gradually eliminated during the first years of the analysis. 24 Figure 2: Gradual Subsidy Removal: Non-Revenue Neutral 2.80 Percent increase in each sector relative to the 2004 2.60 Forestry 2.40 Mining 2.20 Transportatioin Electricity baseline year 2.00 Pet&NG 1.80 Services 1.60 Investment 1.40 Manufacturing 1.20 Refining Agriculture 1.00 Chemicals 0.80 Source: Authors. 33. Consumption goes up in all sectors with the exception of energy and gasoline (because of the large subsidies that both commodities receive in the business-as-usual case and that are eliminated in this scenario). 34. Finally, except for petroleum, chemicals, and refinery products, trade in most commodities is largely unaffected by the elimination of energy subsidies. This is explained by the fact that eliminating subsidies has a local effect, bringing local prices closer to international prices. 35. As noted, the aggregate effect of subsidy elimination would seem to be highly positive, and this is fully in accord with economic theory since resources are directed to more productive areas of the economy. Initially, however, growth goes down slightly for several years. This is due to the fact that there is unemployment in the model, and both labor and capital need a short time to adjust as the initial effects of an energy subsidy removal provide a shock to the economy. Hence, we should not be surprised to see this result, especially since both capital and labor participation grow rapidly throughout the rest of the analysis. Not only does the directly measured level of welfare increase here, but a welfare gain also comes from considerations external to our analysis in the model, such as lower emissions and more sustainable groundwater sources that will be discussed later on. 36. First, and most important, the results in Tables 6 and 7 show that sectors that consume fossil fuels decline substantially from the business-as-usual case. This, in turn, means that the emissions of pollutants such as carbon dioxide (CO2), sulfur dioxide (SO2), nitrous oxides, carbon monoxide (CO), 25 and particulates will also decline, resulting in a large unambiguous increase in welfare through better air quality and, therefore, health. 37. Second, if the elimination of subsidies were to be accompanied by a transfer to the poorer agents, as we will see in Scenario 4, there will be larger equity gains. These environmental co-benefits will be discussed later in the report. Table 7: Change in Consumption, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) Food -0.3580 -0.2945 0.6063 2.1070 Household goods -0.5087 -0.4493 0.5024 2.0484 Consumer services -0.5929 -0.4951 0.4027 1.9066 Autos -0.7860 -0.6573 0.3002 1.8484 Electricity and LPG -8.4621 -12.9987 -11.9806 -10.3041 Public transport -0.6107 -0.7806 0.1122 1.5883 Gasoline -1.8651 -2.8691 -2.1657 -0.4723 Water -0.4622 -0.2571 0.6452 2.1641 Housing -0.5644 -0.4711 0.4300 1.9231 Source: Authors. 38. One more point that is worth mentioning is the relevance of technological change. Even though we do not explicitly model it in any of the simulations, substitution elasticities across fuels are built into the model, so when relative prices of different energy sources change, adjustments in the fuel matrix take place. Technological change could also be modeled explicitly when energy efficiency is modeled as a scenario rather than as a result. Alternatively, some models take technological change as an endogenous variable, reacting to any changes in energy prices. Again, trade in most commodities is largely unaffected by the elimination of energy subsidies. These results are presented in Appendix D20. Caveats 39. Some caveats have to be made with regard to the definition of energy subsidies and the way subsidies are modeled in this exercise. Most of these issues are explained in the footnotes above but it is important to discuss them overall. Subsidies are defined comparing final sales prices with production costs for electricity or considering international prices for other fuels. Occasionally, high subsidies reflect high inefficiencies in production. Subsidies could be reduced significantly by reducing production costs. If subsidies are cut back this way, then there will be effects on welfare if the costs are reduced inexpensively, otherwise there may be a need to factor them into the calculations. 40. We are taking subsidies for 2010 and assuming they remain at the same rate in the business-as usual-scenario. This may be misleading since energy prices, particularly those of oil and gasoline vary in time, and therefore so do subsidies and final prices that may respond to volatility in world markets. This may affect the size of the subsidy since these international prices may be determining the final 20 Trade results for selected scenarios are presented in different parts of the Appendix. Results for other scenarios that are not listed are available from the authors upon request. 26 sales price. In any case we take the subsidies in 2010 as an example, but other exercises could have been made using an average over some period or some other projection for subsidies based on expected changes in production costs and international prices. Sensitivity Analysis – Single-Year Option 41. We did sensitivity analysis to this scenario by simulating a sudden elimination of energy subsidies. The results are very similar to the previous scenarios but with greatest adjustment costs in the initial years. In this case, instead of gradually removing the energy subsidies between 2012 and 2018, we complete the entire policy in a single year. This has the advantage of avoiding delays. It does, however, have the disadvantage of causing a sudden disruption to the economy when it is first done. As before, to concentrate solely on the substitution effects of the subsidy removal, the policy is done in a revenue-neutral manner with the excess government funds being distributed to the lowest two income groups. We compare our results here with those of the business-as-usual case. 42. We see that a sudden removal of energy subsidies increases aggregate welfare slightly. This is a change in the same direction, albeit a bit less than when the subsidies were removed gradually. It would seem, then, that a sudden removal of subsidies is a little more disruptive than removing them gradually. The change, however, is not large. As in Scenario 3, GDP declines slightly in the first year but increases strongly in all subsequent years. Investment again increases throughout the analysis, and the final level of the capital stock goes up. The welfare level of the poorest agents rises as the added government funds are siphoned off to Agents 1 and 2. Agents 3 and 4, again experience more modest percentage gains but definitely gain welfare over the period of the analysis. Economic growth declines in the first year of the analysis but then turns around immediately and then speeds up in the later periods, as before. As with welfare and the level of the capital stock, all of the other aggregate variables show growth relative to business as usual, but the improvement is not quite as pronounced as when subsidies were removed gradually. 43. When we examine the individual sectors, we find that, as in Scenario 3, most sectors increase from the business-as-usual case for the entire period of the analysis. We see that electricity production now goes down compared to the business-as-usual case, almost the same as before. Again, there is a large decline in the petroleum, natural gas, and refinery sectors, while the levels in all other sectors rise. 44. Consumption again goes up in all sectors except in energy and gasoline (again because of the large subsidies that both commodities now receive that are eliminated in this run). Most other sectors see gains, but transportation (another recipient of energy subsidies) sees those gains only in the very last part of the analysis. Food and water experience the largest percentage gains due to the government’s lump-sum adjustments to the poorer agents. Scenario 4. Removal of Energy Subsidies under Revenue Neutrality 45. We have looked at the effect on the Mexican economy of gradually removing energy subsidies. Here, however, the policy is done in a revenue-neutral manner, with all funds from the subsidy removal collected by the government being returned to the various agents so that government funds are 27 the same as in the business-as-usual case. Lump-sum transfers are made to all agents in proportion to the population of that income group so as not to favor either rich or poor consumers. Again, to give us a frame of reference we compare our results with those of the business-as-usual case described in Scenario 2. 46. First, looking at the aggregate numbers, we see that applying energy subsidies increases the aggregate level of welfare by 0.73 percent, which is slightly lower than the non-revenue-neutral case. As before, the numbers are positive for all agents but Agent 4. The reason for this progressive result is that the wealthiest consumers consume a higher percentage of gasoline than their lower-income counterparts and, hence, despite a lump-sum transfer, the level of their welfare slightly decreases. Economic growth declines here until 2018, but becomes positive during the rest of the analysis. Table 8: Change in Aggregate Results, Neutral Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP -1.0231 -0.1000 1.2327 2.0839 Investment -2.6421 0.9031 1.7038 16.8214 Government 0.0000 0.0000 0.0000 0.0000 Capital stock — — — 7.1003 Total welfare (∑Agent 1–4) — — — 0.7252 Agent 1 — — — 6.6970 Agent 2 — — — 2.2141 Agent 3 — — — 0.3186 Agent 4 — — — -0.4779 Government welfare — — — 0.0000 Aggregate welfare — — — 0.7252 Source: Authors. Note: — = Results not reported. 47. One result of all our subsidy scenarios that seems a bit perplexing at first is that the expenditures of all agents go down throughout the analysis when their cumulative welfare goes up. This occurs because consumer spending makes up only part of total economic welfare. In addition to spending their money, consumers get utility and hence welfare out of their leisure time. Furthermore, Agents 3 and 4 have the option of saving their money instead of spending it, and the income that goes into savings also adds to utility. On balance, then, consumer welfare goes up in spite of slightly less spending activity (Table 8)21. Welfare gains are larger for lower income agents because they consume proportionally less energy than the rich and therefore they lose less of their purchasing power once subsidies to energy are removed. Additionally, the fact that Agents 1 and 2 benefit from the lump-sum transfers as opposed to the rich, give them further welfare gains. Overall Agent 4 loses because it is unambiguously affected by the subsidy removal given their large energy consumers. 48. Turning next to the individual sectors, we find that, both within the production sectors (Table 9) and the consumption sectors (Table 10) results are almost exactly the same as in the non-revenue- 21 Leisure goes up when subsidies are eliminated because there is substitution between labor and leisure and the extra income earned allows consumers to work less and increase their utility by having more leisure time. 28 neutral case. Indeed, since the amount of redistributed wealth is quite a bit less than aggregate output, these results are entirely in keeping with our expectations. Furthermore, there is no substantial change in the international sectors, with the exports of fossil fuels and chemicals falling quite sharply and little else changing very much. Table 9: Change in Production, Neutral Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Sector (%) (%) (%) (%) Agriculture -1.4552 -1.0632 1.4015 9.1847 Livestock -1.1851 -0.3432 2.4275 10.3053 Forestry -0.4310 0.3597 3.2738 10.6729 Fisheries -1.3333 -2.2727 -0.9524 2.7451 Oil -5.1397 -8.3000 -6.6805 0.1205 Natural gas -5.1255 -8.2809 -6.6860 0.1684 Mining -3.3189 -2.4390 1.5106 12.7644 Refining -8.7559 -14.8590 -12.5163 -3.8734 Transport -0.7574 -0.8183 0.4732 4.0369 Electricity -24.1957 -35.6632 -33.1683 -25.5993 Chemicals & plastics -4.9750 -5.9092 -1.0854 13.9301 Services -0.3938 -0.0801 1.1363 4.5538 Manufacturing -1.5274 -0.1223 3.7084 13.9623 Source: Authors. Table 10: Change in Consumption, Neutral Subsidy Phaseout vs. Business as Usual 2012 2018 2024 2030 Sector (%) (%) (%) (%) Food 0.4274 0.3483 0.9859 2.1776 Household goods 0.0266 -0.0821 0.6148 1.8463 Consumer services -0.3871 -0.4347 0.2056 1.4071 Autos -0.7860 -0.8381 -0.1256 1.1005 Electricity and LPG -7.7485 -12.4200 -11.7394 -10.3542 Public transport -0.0528 -0.3553 0.2801 1.4488 Gasoline -1.7775 -2.9195 -2.2619 -1.0734 Water 0.3058 0.2558 0.8584 2.0758 Housing -0.3584 -0.4281 0.1796 1.3897 Source: Authors. 49. Finally, as before, fossil fuel use declines substantially from the business-as-usual case, resulting in a large unambiguous increase in welfare reflected as health benefits. Sensitivity Analysis – Different Schemes of Lump-sum Transfers 50. We did sensitivity analysis to see the impact of different schemes of lump-sum transfers. First we look at the case where only 50 percent of the saved resources are transferred to all the population as lump-sum payments. Then we analyze the impact of only giving transfers to the poorest 50 percent of the population. Finally, we discuss the results of simultaneously eliminating energy subsidies and 29 expanding the VAT to food and medicines as a building block toward financing a more ample healthcare system that will be presented in Scenarios 5 and 6. Transfers to the poorest 50 percent. 51. Here we look at the effect of gradually removing energy subsidies and giving back the total savings from the subsidies to the lowest two income groups in a revenue-neutral manner.22 We compare our results with those of the business as usual scenario. Results seem very similar to the original case; GDP declines slightly initially (that is, in 2012) but rebounds quickly in 2018. By the final period of the analysis GDP grows, as does investment and the final capital stock. The aggregate level of government welfare is unchanged (by design), and the welfare level of the poorest agents rises as the added government funds are siphoned off to Agents 1 and 2. Agents 3 and 4, however, see smaller percentage increases. Economic growth declines only in the very earliest part of the analysis and then speeds up in the later periods. However, in this case we see fairly large increases in the agricultural sector as low-income consumers are given transfers and spend a large portion of their disposable income on food and related products. Transfer of 50 percent of saved subsidies. 52. Just as before, we calculate the impact of gradually removing the existing energy subsidies. The only difference from the previous scenario is in the size of the lump-sum transfer. Here, we transfer only half of the funds so that in this scenario we are not revenue neutral. As would be expected, the results are quite similar to the previous scenarios and require very little added explanation. The only thing to note in comparing all subsidy removal scenarios is that the total welfare gain seems to be slightly higher in the case where the government was allowed to keep all of the funds (non-revenue neutral). This, in turn, suggests that a subsidy removal might be most effective when used to finance other programs, as we will see below. The numbers for this scenario are listed in part E of the Appendix. A building block toward more resources. 53. In preparation toward financing a large-impact public policy such as expanded healthcare, we look at the impact of eliminating energy subsidies and at the same time expanding the VAT to food, medicines, and medical services. These health programs are usually expensive and savings arising from the elimination of energy subsidies are most likely not enough to finance such an endeavor. The purpose of this simulation, then, is to see how much revenue can be generated by a combination of these new taxes and the elimination of energy subsidies. In so doing, we also wish to ascertain the impacts of this policy on the level of consumer welfare, fossil-fuel use, and economic growth at both an aggregate and sectoral level. Results are presented in part F of the Appendix as percentage deviations from the business as usual scenario. 54. When we look at the aggregate welfare and GDP numbers, we see that a combination of energy subsidy removal and VAT tax increases leads to gains in the aggregate level of welfare. Now, 22 Revenue neutrality is achieved due to the cumulative effect of the removal rather than on a year-by-year basis. Thus, in our results, we do not necessarily have revenue neutrality in each individual year. Government revenues are, however, neutral over the entire period of the analysis. 30 however, the welfare level rises by only 0.396 percent because a welfare “addingâ€? subsidy removal has been combined with a welfare “takingâ€? tax increase. As before, aggregate GDP declines initially and then goes up. The declines, however, are more severe than in previous cases and the gains are more modest. Furthermore, the losses persist until the very last period of the analysis. 55. Aggregate investment shows a similar pattern for GDP but, as before, the gains in investment are significant in the final years of the analysis, and, as a consequence, the level of the capital stock again rises by over 7 percent. Because of the added taxes, government revenues rise significantly and the cumulative level of government revenue goes up by about 4.36 percent. Cumulative welfare also rises for Agents 1 and 2. Agents 3 and 4, however, experience declines in welfare of 0.25 percent and 0.93 percent, respectively. Hence, we can conclude that a combination of tax increases and subsidy removals leads to higher government revenue totals but at the expense of some consumer welfare and some economic growth. 56. Looking more closely at where the gains and losses occur, we find that the consumption sectors are initially hurt as the policy takes effect, and, only in the very latter stages of the analysis do most of the numbers start rising compared to the business-as-usual case. As before, the gasoline and electricity sectors are the hardest hit. Now, however, serious declines result in foods, consumer services, and consumer goods because the VAT hits these items, as well. Negative impacts are also more persistent in the production sectors. By the final period of the analysis, however, the results return to a fairly familiar pattern, with all sectors experiencing gains except the fossil fuel and electricity sectors. 57. As in previous cases, the international sectors are largely unaffected except with respect to the import and export of fossil fuels, and, as before, less production in the fossil fuel and fossil-fuel-related sectors leads to a decline in the externalities associated with greenhouse gases and other forms of airborne pollution. Scenario 5. Removal of Selected Private Social Security Contributions and Expansion of the VAT: Financing an Expanded Healthcare Package 58. In Scenarios 3 and 4, we were primarily concerned about the energy sector in the Mexican economy and the general equilibrium impact of a removal of the present system of energy subsidies on both aggregate and sectoral variables over time. In Scenarios 5 and 6, the focus of our analysis changes somewhat as we look at the general equilibrium impacts of a more extensive type of tax reform aimed at higher economic growth by increasing the quality of human capital, namely through enhanced access to healthcare. To be more specific, we examine the dynamic implications of an expanded healthcare reform. 59. Based on several authors, we model a fundamental change in the Mexican health policy accompanied by a radical overhaul of the present tax system and changes to the private contributions to social security. The main modification to the tax system that we include deals with having food and medicines pay the same 16% rate of the VAT that all other goods pay. This is an ongoing discussion taking place in Mexico that at this stage could be promoted given that it may generate additional resources to pay for the expansion of the healthcare system 23. This would have the effect of moving 23 Obviously there are other loopholes that would have to be closed to make the tax system more efficient, but at this point this is out of the scope of this paper. 31 labor from the informal sector (which does not formally and consistently contribute to social security) to the more efficient formal sector (which is currently taxed to provide security and health benefits). 60. To make up for the loss in revenue from these contributions, we recommend extending the current VAT (which is 16 percent) to food, medicines, and medical goods and services that are currently exempt. In addition, a certain amount of all VAT revenues would be earmarked to cover the new national healthcare and social security system, and the government would subsidize the remainder from energy subsidies. This to some extent follows Anton et al 2012. As a note of caution, even though we think theirs is an interesting proposal, we do not model it exactly as they present it. For example, we do not eliminate private employee-employer contributions toward daycare, housing, and retirement savings. We do not compensate for the extra VAT that lower-income groups are paying. Nor do we model any of the extensions they present in their paper. We do take some of their numbers, such as the overall cost of Mex$560 billion a year as an estimated cost of a comprehensive healthcare program24. Other authors believe this number is too low for such a program, but we use this amount only as a ballpark estimate of possible costs, acknowledging that there may be infinite combinations of costs and policies. To see the main details of Anton et al 2012, see Appendix G. 61. As before, to measure the impact of the parametric changes in this simulation, we compare our results to those of the business-as-usual case and give the percentage changes. Overall, the changes turn out to be dramatic and quite welfare enhancing. Furthermore, all the numbers tend to increase over time, lending support to the notion that this kind of policy could be quite effective for the Mexican economy. 62. Turning first, to the aggregate impacts, we see that there is a large increase in GDP over time. Indeed, when compared to business as usual, the level of GDP grows by about 2.7 percent in 2018 to about 2.8 percent in 2030. These increases are primarily due to two factors. 63. First, as some of the private social security contributions are removed and labor moves from the informal to the formal sector, more labor is hired. The level of unemployment goes from over 3 percent in the business-as-usual case to no unemployment in the latter years of this analysis. 64. Second, following the stimulus to the aggregate economy, the level of investment increases substantially because the government has additional resources to spend, that come from the VAT to food and medicines, and labor is moving towards the formal sector, which is more productive and more capital intensive. This jump in investment is most apparent in the early years, when it grows by more than 7 percent over the business-as-usual case. The increase declines a bit in later years. Still, it remains above the business-as-usual case throughout the time of the analysis. Spurred by the increase in investment, the final level of the capital stock rises by over 7.9 percent. Furthermore, the welfare level of all agents rises, with the changes being slightly progressive in nature, as seen in Table 11. This occurs to some extent because of the transfers given to the poorest consumers. Still, some of it is due 24 Anton et al 2012 assume that the current cost in health worker is $10,118 pesos. Because of a work force of 39.03 million workers, the cost of health insurance would be $394.9 billion pesos (mmp). It should be added to this amount $136.3 mmp to cover pensions for retirement and $28.1 mmp for life and disability insurance. This gives a total cost of this package of $559.3 mmp according to Anton, Hernandez and Levy (2012). We take this number for our computations as one possible example, given this is the most clearly described in a published document. 32 simply to the nature of the changes that are designed not to inordinately enhance the welfare of the higher-income agents. Table 11: Change in Aggregate Results, Expanded Healthcare vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP 1.2342 2.6968 2.7954 2.8201 Investment 7.1014 5.6811 4.4193 4.3329 Government -3.8546 -1.1842 -0.5012 -0.2325 Capital stock — — — 7.9196 Total welfare (∑Agent 1–4) — — — 4.6505 Agent 1 — — — 5.1138 Agent 2 — — — 4.9974 Agent 3 — — — 4.5227 Agent 4 — — — 4.5297 Government welfare — — — -3.7730 Aggregate welfare — — — 3.2045 Source: Authors. Note: — = results not reported. 65. Finally, the tax changes are not revenue neutral and total government income does decrease over the period of the analysis. It bears noting, however, that in the later periods of our simulation, the level of consumer and producer income rises and government revenue is only slightly lower than in the business-as-usual case (though deficits are run throughout the analysis). 66. Because of the complex links among sectors in our model and the multifaceted nature of the tax changes implemented, these changes are a bit harder to interpret than the aggregate changes listed above. Nonetheless, the relative magnitude of these changes can be understood if we look at the basic ways in which the tax reform proposal affects supply and demand in our model. 67. First, as we have noted, the decrease in some of the labor taxes (namely, selected social- security-related contributions) is accompanied by a corresponding increase in labor supply. Hence, the most labor-intensive sectors will be the principal beneficiaries of this supply shock. Second, directly following the change in tax policy, there will be an increase in investment as the income of investors (that is, savers) increases, and this increases the demand to those sectors that are most responsible for the production of investment goods. Third, because of the expansion of the VAT, food and consumer goods (where medical services are accounted for) receive the largest negative shock. Fourth, the income of the government goes down, and this reduces the demand for goods and services (that is, mainly from the services sector). Finally, all production sectors are connected due to input-output effects, and if demand goes up in one of those sectors, the demand will also rise in those sectors that supply most of the inputs. The opposite, of course, happens in the case of a demand decrease for a particular sector. 68. With these general concepts in mind, we now look at the increases for the various production and consumption sectors. In our model, the highest capital-to labor ratios occur in the chemical and manufacturing sectors, while the lowest occur in the agriculture, livestock, and forestry industries. 33 Furthermore, the input-output matrix used in our model reveals that the highest investment demand (as a percentage of total demand) occurs in the manufacturing, chemicals, petroleum and natural gas, and refinery sectors, while the lowest investment demand occurs in services, fisheries, and transportation. Hence, we see that the production sectors experiencing the biggest increases include chemicals, petroleum and natural gas, refineries, and manufacturing. Mining also rises substantially due to the fact that most of its inputs go into manufacturing, refining, and chemicals. 69. Conversely, we see that livestock, agriculture, fisheries, forestry, and transportation experience substantial but lower gains. Part of the reason for the smaller gains in the agricultural sector is also likely due to the heavier taxes on food at the consumer level. Finally, we observe that the smallest percentage gains are experienced by the service sector. However, it must be remembered that Table 12 gives only percentage changes, and since the services sector is the largest production sector in Mexico, its “totalâ€? gains are approximately the same as those for manufacturing. Table 12: Change in Production, Expanded Healthcare vs. Business as Usual 2012 2018 2024 2030 Sector (%) (%) (%) (%) Agriculture 0.2202 4.7731 7.9700 9.6781 Livestock 2.6316 6.1363 8.6482 9.9975 Forestry 2.1008 5.7823 8.9636 9.4651 Fisheries 1.9355 5.2632 7.4236 9.8182 Oil 1.4500 7.0386 10.8557 13.7680 Natural gas 1.4706 7.1043 10.8502 13.8081 Mining 4.6605 9.0909 12.1294 13.8702 Refining 2.8674 7.2950 10.4528 12.8005 Transport 1.9002 5.2016 7.5256 9.5329 Electricity 3.4859 6.6087 8.8356 10.9830 Chemicals & plastics 5.4406 11.3406 15.7309 18.5830 Services 0.3851 3.6823 6.0085 8.0639 Manufacturing 4.6090 8.3096 10.9359 12.2901 Source: Authors. 70. Turning now to the consumer goods sectors, we see that goods and autos, with their high manufacturing and chemical content, stand to gain the most. Food experiences a slight increase by 2030 in its final consumption due to reasons outlined at length above. Interestingly, gasoline ranks slightly lower than all consumer goods in terms of its demand increase over the course of the simulation, even before eliminating energy subsidies, as is the case here in table 13. 34 Table 13: Change in Consumption, Expanded Healthcare vs. Business as Usual Category 2012 2018 2024 2030 Food 1.0698 4.1242 6.2781 8.5872 Household goods 2.5994 5.4216 7.5023 9.8051 Consumer services 2.2033 5.1237 7.1518 9.6404 Autos 2.4550 5.3917 7.4762 9.8051 Electricity and LPG 2.1898 5.2859 7.4487 9.7661 Public transport 2.1689 5.2024 7.3399 9.6057 Gasoline 2.0690 5.1549 7.3162 9.6045 Water 2.2489 5.2248 7.4148 9.7338 Housing 2.1956 5.2121 7.3358 9.6095 Source: Authors. 71. Finally, since the changes modeled here are primarily domestic in nature, there is little change in the foreign trade of goods and services. Both imports and exports go up as domestic consumption and production increase, but all changes there are relatively minor. Sensitivity Analysis – Variations in Elasticity of Substitution across Formal and Informal Labor 72. To check the robustness of our results, we vary the elasticity of substitution between informal and formal labor in each of our production sectors. Initially, we set that elasticity at 1 (that is, the Cobb-Douglas case) and then expand it to see the impact such a change has on our models variables over time. We then increase the elasticity of substitution between informal and formal labor in each production sector from σ = 1 to σ = 2. This allows labor to travel more easily from the informal to the formal sector following a tax reduction, and allows the economy to adjust more quickly and efficiently to the tax change.25 Given the relative ease with which labor can be substituted between the formal and the informal sector, this is probably a fairly realistic assumption. The results produced in this simulation are highly similar to those of our previous case. 73. We finally look at the case where the elasticity of substitution between informal labor and formal labor in all production sectors is increased from 2 to 3. These sensitivity analyses are important because, along with the level of economic growth that we assume, the elasticity between labor in the informal and formal economies is the most important parameter to be considered. Furthermore, since it is difficult to find a reliable estimate of this parameter, it is critical that the comparative statics and dynamics in our study be robust with respect to parametric changes here. 74. Basically, what our model shows under higher rates of substitution between formal and informal labor is that an increase in the substitution of elasticity from 2 to 3 has a very similar effect to when it was increased from 1 to 2. Again, the aggregate numbers go up slightly as labor becomes more substitutable. The final level of the capital stock increases. Growth is also seen in aggregate investment, GDP, and government spending in all periods following the healthcare and tax reform. 25 More formally, the elasticity σ here represents the change in the percentage quantity of labor (in each sector), which goes from the informal to the formal sector following a 1 percent lowering of the relative price of labor in the formal sector. Such elasticities exist between all sets of inputs, including labor and capital. In the MPSGE (Mathematical Programming System for General Equilibrium) model that we use, the production in each sector follows a “nestedâ€? CES structure, allowing us to change these elasticities in the model with relative ease. 35 Aggregate welfare for all agents goes up over the previous case, with the largest gains accruing to the poorer agents. 75. Welfare gains for all income groups tend to be marginally close to zero, indicating that, as before, the gains due to larger substitutability are positive but not particularly large. Welfare gains for all groups, however, go up over time with the accumulation of more capital stock. As before, consumption gains are higher in food, leading to slightly larger increases in the agriculture and agricultural-related production sectors. Gasoline consumption also rises a bit more than the average. Foreign sectors are largely unaffected, as in all previous cases. 76. Finally, it should be noted that model convergence was the most difficult in this particular case of higher substitutability. The relatively high level of the elasticity between labor types caused some concavity issues, and it required more iterations to achieve tangencies between the isocost and isoquant surfaces in the nested production functions used by the model. Scenario 6. Dual Policy Implementation 77. In this final scenario, we quantify the economy-wide and sector-specific impacts of an expanded healthcare system financed by a VAT expansion and coupled with energy subsidy removal and a reduction in some of the private contributions to social security. Namely, this is a combination of Scenarios 3 and 5, that is, the non-revenue-neutral energy subsidy removal scenario and the expanded healthcare reform scenario presented above. We list the percentage changes of this scenario from the business-as-usual case. This allows us to see the total impact of carrying out these policy changes, and as can readily be seen from there, this is basically a combination of the two. It is important to remember when examining these impacts, however, that this is not a linear model and that the changes seen here are not strictly a linear combination of the two previous policy change scenarios. 78. The increase in the level of the aggregate variables brought about by dual policy implementation is both substantial and persistent. As we can see, total welfare in this case rises by 4.2 percent over the business-as-usual case. This represents a fairly substantial case and is fully 1 percent greater than the expanded healthcare and VAT policy results we discussed in Scenario 5. Significantly, the government deficit (relative to the business-as-usual case) is only about .07 percent over the entire period. This, in turn, suggests that the expanded healthcare policy could be almost totally financed with the help of revenues gained from subsidy removal and expansion of the VAT. Both GDP and investment go up in all periods from 2012 on, and by 2030 the level of GDP has risen by about 3 percent over the business-as-usual case. The increases in investment lead to a 7.5 percent increase in the capital stock, and the subsidy removal combined with a growth in income leads to a growth in government revenue by the latter part of the analysis Table 14). 36 Table 14: Change in Aggregate Results, Joint Policies vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP 0.7517 1.9120 2.8208 3.0290 Investment 6.0117 4.6914 4.4717 4.4506 Government -0.2063 2.4550 3.1721 3.4827 Capital Stock — — — 7.5246 Total welfare (∑Agent 1–4) — — — 3.6562 Agent 1 — — — 3.8493 Agent 2 — — — 3.7452 Agent 3 — — — 3.5033 Agent 4 — — — 3.6871 Government welfare — — — -0.0736 Aggregate welfare — — — 4.2177 Source: Authors. Note: — = results not reported. 79. Turning now to the results in the consumption sectors, we see that the aggregate impact of dual policy implementation is quite encouraging. All sectors with the exception of energy see gains. The loss in energy consumption occurs as a consequence of subsidy removal, and is a good thing from a policy standpoint, since a cleaner environment is a goal of this particular simulation exercise. Overall gasoline consumption goes up by about 4.6 percent by 2030, since the marginal effect of subsidy removal is more than offset by the effects of the VAT tax reform (Table 15). Nonetheless, gasoline consumption grows by only half as much as when there was no subsidy removal (see Table 13). Table 15: Change in Consumption, Joint Policies vs. Business as Usual 2012 2018 2024 2030 Category (%) (%) (%) (%) Food 0.0167 2.7290 5.0913 7.5731 Household goods 1.4182 4.0003 6.3043 8.7493 Consumer services 1.3101 4.0216 6.3465 8.7971 Autos 1.6045 4.2354 6.5432 9.0148 Electricity and LPG -6.7343 -8.2614 -5.6517 -2.8354 Public transport 0.9153 3.3576 5.7203 8.1597 Gasoline -0.0564 1.6643 4.0685 4.5864 Water 1.2121 3.9409 6.2880 8.7468 Housing 1.2594 3.9705 6.3048 8.7503 Source: Authors. 37 Table 16: Change in Production, Joint Policies vs. Business as Usual 2012 2018 2024 2030 Sector (%) (%) (%) (%) Agriculture -1.4931 2.4744 6.0584 8.2192 Livestock 1.3333 4.6394 7.5844 9.2414 Forestry 1.2712 5.4608 8.7079 10.6729 Fisheries 0.0000 1.6393 4.0724 6.4151 Oil -3.0570 -0.5104 3.3912 6.3818 Natural gas -3.0769 -0.4864 3.4211 6.3931 Mining 1.9178 5.7239 9.2764 11.2903 Refining -5.0018 -5.9706 -2.2452 0.3924 Transport 0.5081 3.1979 5.7540 7.9797 Electricity -19.6943 -26.9004 -23.5970 -20.5423 Chemicals & plastics 0.9265 4.6952 9.5167 12.4518 Services 0.1908 3.2893 5.7892 8.0312 Manufacturing 3.4971 7.1829 10.4092 12.0129 Source: Authors. 80. The results in the production sectors are similar to those in the consumption sectors in that they show the overwhelming impact of tax reform of increasing production and consumption almost everywhere. Table 16 shows that all sectors experience gains except the electricity sector. Electricity is a big recipient of the current subsidies, and the marginal impact of subsidy removal is most severely felt in that sector. Refining, natural gas, and petroleum increases are all significantly less than in Table 12 because subsidy removal cuts back on their growth. Nonetheless, with the exception of refining, all are positive for most of the period of the analysis. 81. Figure 3 shows how different sectors grow after both policies have been implemented. Compared to Figure 2, which shows the non-revenue-neutral gradual elimination of energy subsidies, the trends of growth across sectors is fairly similar now under the joint policy. Investment does not grow as much as in Figure 2 since it has now been used to finance expanded healthcare. The energy- related sectors behave quite similarly under the two scenarios. 82. Finally, since all policy changes considered here are concerned with domestic policy, there is little to report with respect to the foreign sector (trade effects are listed in Appendix H). With that said, however, we should note that there is an overall increase in both imports and exports. More domestic production and higher consumption spawns foreign trade activity and increases trade links with foreign trade partners. 38 Figure 3: Eliminating Energy Subsidies and Financing Universal Healthcare 2.60 Percent increase in each sector relative to the 2004 2.40 2.20 Forestry Mining 2.00 Transportatioin baseline year 1.80 Electricity 1.60 Pet&NG Services 1.40 Investment 1.20 Manufacturing 1.00 Refining Agriculture 0.80 Chemicals Year Source: Authors. 3. Environmental Impacts of Reforms 83. One of the goals of this study is to address the environmental effects of eliminating energy subsidies. Lower emissions are a goal in itself for the Mexican economy, which is now looking at low emission development strategies that focus significantly on mitigation. In addition, it has implemented the Special Program on Climate Change (Programa Especial de Cambio Climático, PECC) to achieve this. Eliminating subsidies to energy use may contribute significantly toward a low-carbon economy, as can be seen from this exercise. 84. Eliminating subsides to energy use across the board leads to lower use of fossil fuels and thus to a reduction in emissions (a list of country-level emissions proportional to GDP is shown in Appendix I). In this analysis, we look at only CO2 emissions from the burning of fossil fuels. We do not consider any other greenhouse gas emissions or any CO2 emissions from land use change. 85. Looking at energy use, we estimate that 41.7 million tons of CO2 are abated every year during the period of analysis. PECC has an abatement goal of 51 million tons per year for 2012. This implies that only by eliminating energy subsidies are we able to meet roughly 80 percent of Mexico’s annual emission abatement goals. As mentioned above, this does not include other greenhouse gases that 39 could add to even more CO2e abatement. These results are encouraging since they highlight the environmental benefit of these types of policies. In addition, they uphold the argument that elimination of energy subsidies is not regressive in terms of income distribution. 86. Lower CO2 emissions also imply lower local emissions, leading to improved environmental quality across the board. This is relevant since a rough but educated guess shows that air pollution imposes a yearly cost of 1.5 percent of GDP on Mexico. This only measures the urban impacts on health in Ciudad Juárez, Guadalajara, León, Mexicali, Mexico City, Monterrey, Puebla, Tijuana, and Toluca, as estimated by the World Bank for 2012 (Table 17). Table 17: Burden of Health Impacts of Urban Air Pollution Health Impact Cases Million US$ Premature mortality 12,220 123.94 Premature mortality, children 1,934 19.62 Chronic bronchitis 19,648 1.79 Hospital admissions 56,760 2.81 Emergency room visits 1,113,442 4.78 Restricted activity days 187,662,750 54.19 Lower respiratory illness in children 2,478,047 4.82 Respiratory symptoms 597,257,100 3.66 Total 216 (1.5% GDP) Source: Authors. 87. Finally, these types of policies are bound to have an impact on other natural resources 26. For example, currently there is a subsidy to electricity used in water pumping for agriculture. This has led to a significant overexploitation of groundwater resources. Policies such as these that increase the effective cost of water extraction will lead to a more rational use of water and therefore less overexploitation of the resource (Ã?vila et al 2005). 4. Policy Implications and Future Work 88. The use of a computable general equilibrium model of the Mexican economy allowed us to analyze the economic, distributional, and environmental impact of energy subsidy reductions. When subsidies to energy use are eliminated, either gradually throughout 2012 to 2018 or suddenly in 2012, distortions are reduced and government resources are liberated, but short-run negative welfare effects take place. To avoid this we test different schemes to transfer those saved resources to the population, sometimes to all, others to the lower 50% in terms of income, and sometimes only part of the resources saved from the subsidies. This in itself avoids a stagnation of economic activity and keeps up both expenditure and demand. 89. The energy-intensive sectors are significantly affected because they do not now receive the subsidy. However, since subsidies cause distortions, when they are eliminated there are positive and progressive welfare gains for most income groups, with higher gains to lower-income agents. Effects are quite similar under a gradual and a sudden removal of subsidies, given the same compensation 26 This does not stem from the analysis but is it worth discussing as an effect of elimination of electricity subsidies. 40 mechanisms. It remains to be seen, however, whether effective mechanisms that guarantee lump-sum transfers reach the targeted population. 90. We then simulate the effect of an expanded healthcare program. We run this policy with the model and make some sensitivity analysis for different rates of substitution among formal and informal workers across sectors. The main results in this case are a significant increase in GDP, investment and capital accumulation, significant and progressive gains in welfare, and basically an increase in production and consumption across the board. The general equilibrium effects of the healthcare program, to increase coverage to all workers, whether formal or informal, and to promote formality in the Mexican economy, does indeed lead to a much more productive economy, with a more equitable income distribution, and with lower emissions as well. 91. Finally, when energy subsidies are eliminated and other important reforms such as an expanded healthcare program covering all workers are combined, the results are highly desirable. This leads to the conclusion that these reforms help reduce distortions to the economy; increase its productivity and have positive effects on growth, investment, and capital accumulation; they promote a more equitable income distribution; and reduce emissions. 92. In terms of the environmental effects, eliminating subsidies to energy in itself achieves several goals. It cuts back on the use of a non-renewable resource that is currently facing depletion and the extraction costs of which are rising, and it promotes energy savings and efficiency. This, in turn, reduces emissions that have both global and local effects, contributing to improved air quality and reduced health costs. In addition, this contributes to mitigating climate change, since subsidy cutbacks alone get 80 percent of Mexico’s yearly abatement goals for CO2 emissions, as set by the 2008–2012 Special Program on Climate Change. Finally, since a significant part of energy subsidies goes to pumping water for agriculture, eliminating them may also help reduce groundwater extraction and therefore the replenishment of such sources. 93. As explained above, energy subsidies impose high fiscal costs and distortions to the Mexican economy. Healthcare is not readily available to all workers, and those that are not covered by any of the existing health systems face catastrophic expenditures and expensive private services if they happen to need medical attention. This exercise demonstrates that there may be other productive uses for those energy subsidies that may be welfare enhancing, promote overall economic growth, and that may certainly increase the quality of health of many more Mexicans, both directly and indirectly through a cleaner and more sustainable environment. Future Work 94. As usual, the scope of this paper is limited and throughout the analysis and discussion with colleagues an important set of possible extensions came up. Some recommendations came up regarding what could be built into the baseline, what those resources saved from energy subsidies could be used for, and finally other interesting simulations that could provide resources. 95. In terms of the baseline, for example, if Mexico decides to exploit its natural gas reserves of tight gas, the baseline would change and this in turn could change relative prices of fuels. Assumptions 41 on a slower or faster rate of oil exhaustion could also be built into the baseline and that would lead to a different set of possible simulations. 96. As for the type of programs that could be financed with resources from energy subsidies, the conditional cash transfer program Oportunidades could be expanded, investment in energy efficiency equipment that in the long run will translate into energy savings, generation of clean energy, and to the provision of cleaner and more efficient public transport. These exercises could definitely be done with the model in the future. Some of these policies have previously been done with the model. 97. Finally, to gain further insight of the energy sector, simulations to address the marginal effect of eliminating subsidies to each energy source could be done separately. This would provide a better understanding on the impact economic and distributional effects of different subsidies. Conversely, a set of simulations could be done to see the effects of imposing carbon taxes. This could provide arguments for international negotiations, as well as fund some of the other projects such as clean energy, energy efficiency, and clean transport discussed above. 42 References Antón, A., F. Hernández, and S. Levy. 2012. The End of Informality in Mexico? Fiscal Reform for Universal Social Insurance. Inter-American Development Bank, Washington, DC. Armington, P. 1969. “A Theory of Demand for Products Distinguished by Place of Production.â€? IMF Staff Paper 16, International Monetary Fund, Washington, DC. Atkinson, A. B., and J. B. Stiglitz. 1980. Lectures on Public Economics. London and New York: McGraw-Hill Book Co. Ã?vila, S., C. Muñoz, L. Jaramillo, A. Martínez. 2005. Un análisis del subsidio a la tarifa 09. Gaceta Ecológica. Abril-junio, 75. Instituto Nacional de Ecología, México: 65-76. Ballard, C., D. Fullerton, J. Shoven, and J. Whalley. 1985. A General Equilibrium Model for Tax Policy Evaluation. Chicago: University of Chicago Press. Berndt, R. E. and R. Samaniego. 1983. “Residential Electricity Demand in Mexico; A Model Distinguishing Access from Consumption.â€? Working Paper, Massachusetts Institute of Technology, Cambridge, MA. Bernstein, M. A., and J. Griffin. 2006. “Regional Differences on the Price–Elasticity of Demand for Energy.â€? National Renewable Energy Laboratory, Santa Monica, California. Chávez Presa, J. A., F. Hernández Trillo, and L. F. López-Calva. 2012. “El México del 2012: Reformas a la Hacienda Pública y al Sistema de Protección Social.â€? Centro de Estudios Espinosa Yglesias, México D.F. Charles, C. and P. Wooders. 2011. Subsidies to Liquid Transport Fuels: A comparative review of estimates. The International Institute for Sustainable Development. Geneva. CICC (Comisión Intersecretarial de Cambio Climático). 2009. Programa Especial de Cambio Climático 2009–2012. DOF 28/08/2009. México DF. Fernández, O. 1997. “Efectos de la Aplicación, de un Impuesto Ecológico Neutral en México: Análisis Mediante un Modelo de Equilibrio General Computable.â€? In Instrumentos Económicos para un Comportamiento Empresarial Favorable al Ambiente en México, ed. A. M. García. Mexico City: Colegio de México, pp. 105–19. Ibarrarán, M. E., and R. G. Boyd. 2006. Hacia el Futuro: Energy, Economics and the Environment in 21st Century Mexico. New York: Springer. IEA (International Energy Administration). 2011a. World Energy Outlook 2011, U.S. Department of Energy and U.S. Energy Information Administration. IEA (International Energy Administration), 2011b. CO2 Emissions from Fuel Combustion (2011 Edition), IEA, Paris. Ramsey, F. 1928. “A Mathematical Theory of Saving.â€? Economic Journal 38 (152): 543–59. Reyes Martínez, O., R. Escalante, and A. Matas. 2010. “La demanda de gasolinas en México : efectos y alternativas ante el cambio climático.â€? Economía: teoría y práctica 32.2010: 83–111. México, DF, ISSN 0188-8250, ZDB-ID 14333053. Romero, J. 1994. “Energía, Emisiones y Precios Relativos.â€? In Medio Ambiente: Problemas y Soluciones, ed. A. Yúnez-Naude. Mexico City: Colegio de Mexico, 41–52. SENER (Secretaría de Energía de México). 2012. “Estrategia Nacional de Energía 2012–2026.â€? México DF. Serra-Puche, J. 1984. “A General Equilibrium Model for the Mexican Economy.â€? In Applied General Equilibrium Analysis, ed. H. E. Scarf and J. B. Shoven. Cambridge, England: Cambridge University press, 64–87. Wylie, P. J. 1995. “Partial Equilibrium Estimates of Manufacture, Trade Creation, and Diversion due to NAFTA.â€? North American Journal of Economics and Finance 6 (1): 65–84. 43 Appendix A. The model and recent applications Production Production in each sector for every time period is represented as a constant elasticity of substitution (CES) value added function of capital, labor, energy and material inputs, where the elasticity of substitution can vary between zero and infinity. Hence, ï?³ -1)/ï?³ ( ï?³/ï?³ -1) V t =ï?¦t [ ï?¤ L Lt ( ï?³ -1)/ï?³ ï?³ -1)/ï?³ ï?³ -1)/ï?³ (1) +ï?¤ K K( t  ï?¤ E E( t  ï?¤ M M( t ] where Vt is the value added at time t, σ is the elasticity of substitution between inputs, ï?¦ t is an efficiency parameter that shifts the whole production function, Lt is labor at time t, Kt is capital at time t, Et is energy at time t, Mt are material inputs at time t, ï?¤ are share parameters defined so that δL, ï?¤K, ï?¤M, ï?¤E > 0 y δL + ï?¤K + ï?¤M + ï?¤E = 1 Prices of all goods and services are equal to one, so that Vt refers to quantity as well as value of production. Material inputs are all inputs from the sectors of production, and M t refers to a composite good based on a nested CES production function, where inputs from all sectors are included (See Figure A.1) . These nested functions are used for production and consumption, permitting different levels of substitution between inputs or goods and services. For the production function, substitution is allowed between labor, capital, and energy as well as between energy inputs and non-energy inputs. Figure A.1 Nested production function Source: Ibarrarán y Boyd, 2006. Consumption Total utility for each household c is modeled by (2) Uc = ï?“t Uc,t (Xc,t, Rc,t) * (1+ï?² )-t t = 1, …, n where Uc is household utility over all n time periods, Uc,t is the utility derived from the present period consumption of goods and services, Xc,t (a 7-dimensional vector) and leisure Rc,t, and where ï?² is the rate of time preference. Each Uc is taken to be a nested CES utility function defined for all consumer goods as well as all time periods. The value of household utility is given by the addition of the value of consumption plus the value of leisure, which is equal to the number of hours devoted to leisure times the net wage per hour worked (Ballard et al., 1985). Each consumer’s expenditure constraint can be written as: n (3)  (TG t 1 c,t + TFc,t + (PL,t * Lc,t) + (r * K t * Sc,t)) = n  ((INV * S t 1 t c,t) + (PI,t * Xc,t) + (PL,t * Rc,t)) Where endowments are given on the left-hand side of the equation and expenditures are placed on the right hand side. TGc,t and TFc,t represent the transfer to the consumer from the government and the foreign agents; PL,t is the price of labor y r is 44 the rental of capital. Kt is the level of stock capital in period t; Sc,t is the share of total capital owned by consumer c, INVt is the total investment in time period t; and PI,t is the vector of prices for consumer goods Maximizing the nested utility function (2) with respect to the expenditure constraint (3), simultaneously determines the consumption level of the consumer goods and services, the amount of labor supply, and the consumers level of saving and investment in each of the periods. Government The government sector is treated as a separate agent (Ballard et al 1985). The government agent is modeled with an expenditure function similar to the household expenditure functions (based on a CES utility function). The equations that describe the behavior of the government are: ï?¡ ï?¡ ï?¡ ï?¡n Gu  x1 1 x2 2 xi i xn (4) ï?¡i i 1 n E   Pi ï?¡i i 1 where Gu is the utility function of the government and ï?¡i represents the share of the sectors, and xi are the units that the government consumes. E is total expenditure by the government; and P i are the market prices of the goods and services. Trade International trade within the model is handled by means of a foreing agent. The balance of trade relationship is given by (5) ï?“(Pm,t * IMj,t ) = ï?“(Pj,t * EXj,t ) + ï?“TFc,t t = 1, ..., n where IMj,t is a vector representing the quantity of each to the producer goods imported; Pm,t is the vector of imported goods prices, EXj,t is the vector of producer goods exported, Pj,t is the vector of producer goods prices, including tariffs, and TFc,t is the level of transfers. Labor growth and capital formation The growth of labor over time is given by (6) Lt+1 = Lt(1+ï?§) where ï?§ is the labor growth rate over time. In absence of any perturbation the Ramsey model predicts that the economy will grow at the labor growth rate in the steady state. Capital growth rate is represented by a system of three equations: (7) PA,t = Pk,t+1 t = 1, ..., T where PA,t is the weighted price of consumption and Pk,t+1 is next year’s price of capital. We also have (8) Pk,t = (1+rt) Pk,t+1 t = 1, ..., T meaning that the price of capital in this period Pk,t must be equal to the present period’s rental value of capital plus next period’s price of capital Pk,t+1. Finally we have (9) Kt+1 = Kt(1-Δ) + INVt t = 1, ..., T where Δ stands for the rate of depreciation and INV stands for gross investment. This states that the capital stock in the next period must be equal to this year’s capital stock plus net investment. Taken together, eqs. 7 -9 insure that economic growth will be consistent with profit maximization behavior on the part of investors. Terminal conditions A few adjustments are necessary to design a model which when solved over a finite horizon approximates infinite horizon choices (Lau, Puhlke, and Rutherford, 1997). We divide the problem into two distinct sub-problems, one defined over the finite period from t = 0 to t = T, and the second the infinite period from t = T+1 to T = ï‚¥. Hence, the first problem is T 1 t (10) Max ï?“ ( ) U c,t ( X c,t , Rc,t ) t=0 1+ ï?² subject to T T (11) P t=0 A,t X c,t =  P L,t L c,t + Pk ,0 K c,0 S C ,t  Pk,T +1 K c,T +1 S C ,T +1 t=0 and 45 (11a) Lc,t = Lc,t + Rc,t for all t = 0, 1, ... T And the second problem is ï‚¥ 1  ( 1+ ï?² ) U t (12) Max ct ( X c,t , Rc,t ) t=T +1 subject to ï‚¥ ï‚¥ (13)  t=T +1 PI,,t . X c,t = P t=T +1 L,t Lc,t + PK,T+1 K c,T+1 S c,t+1 (13a) Lc,t = Lc,t + Rc,t for all t=T+1, ...∞ where Ï? is the rate or time preferences, ro and Kc,o refer to the rental value of capital and quantity of capital before the terminal period, rT+1 y K c,T+1 refer to these variables after the terminal period, and L c,t is total labor plus leisure for each agent in the tth time period. PK,t is the price of capital, and PI,t and PL,t are the prices of the consumption goods and the price of labor, both net of taxes. We include the level of post-terminal capital as a variable and add a constraint on investment growth in the final period (14) INVT/INVT-1 = YT/YT-1 where YT gives the GDP at time T. Table A1: Recent applications of the model Study Publication/Institution Cost of Compliance with the Kyoto Protocol: A Developing Paper published in Energy Economics, Country´s Perspective 2002 Hacia el Futuro: Energy, Economics and the Environment in Book by Springer, 2006 21st Century Mexico Macroeconomic Impacts of Mitigation Policies in Selected Project for Centro Mario Molina, 2008 Sectors Evaluación de las implicaciones económicas de las políticas de Project for Instituto Nacional de Ecología, adaptación y mitigación frente al cambio climático en México: 2008 Análisis de equilibrio general de la vulnerabilidad intersectorial Impactos de Cambios en Precios Relativos de Combustibles y de Project for UNAM, 2009 Automóviles MEDEC policies under a general equilibrium approach, Mexico, Project for The World Bank, 2009 2009 Extreme Climate Events and Adaptation: An exploratory Paper published in Environmental and analysis of drought in Mexico Development Economics, 2009 La ruta de México hacia una economía sustentable de baja intensidad de carbono Project for Centro Mario Molina, 2010 Evaluación macroeconómica de políticas de mitigación de gases Project for Programa de las Naciones de efecto invernadero: Hacia una estrategia de crecimiento con Unidas para el Desarrollo e Instituto bajas emisiones de carbono (LEDS) Nacional de Ecología, 2012 Impactos macroeconómicos del cambio climático sobre Project for Programa de las Naciones los principales sectores económicos en México Unidas para el Desarrollo e Instituto Nacional de Ecología, 2012 Source: Authors. 46 B. Building in the Stylized Facts: Business as Usual vs. Benchmark Table B1: Change in Aggregate Results, Business as Usual vs. Benchmark 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP -0.4328 -1.7853 -3.7682 -5.5366 Investment -0.3915 -3.9731 -6.7126 -4.7218 Government 3.5903 3.1886 1.7197 -0.6611 Capital stock — — — -4.3383 Aggregate welfare (∑Agent 1–4) — — — -2.9707 Agent 1 — — — -3.0514 Agent 2 — — — -2.9288 Agent 3 — — — -2.8299 Agent 4 — — — -3.0602 Source: Authors. Note: — = results not reported. Table B2: Change in Production, Business as Usual vs. Benchmark Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 0.0368 -1.2995 -3.8168 -5.8142 Livestock -0.1351 -1.7864 -4.2326 -5.7991 Forestry 0.0000 -1.4440 -4.3077 -5.9740 Fisheries -1.3158 -3.3333 -5.1887 -8.4677 Oil -4.6749 -22.6454 -38.6782 -50.0171 Natural gas -4.6766 -22.6525 -38.6921 -50.0590 Mining 0.4190 -2.2619 -5.8282 -7.9654 Refining -4.0590 -17.6267 -30.0491 -42.3845 Transport -0.8673 -2.3458 -4.6582 -7.3340 Electricity -0.9850 -3.4761 -6.6383 -9.9068 Chemicals & plastics -3.4237 -16.6345 -30.3414 -43.4230 Services 0.0585 -0.8640 -2.7657 -5.0673 Manufacturing 0.2911 -2.1048 -5.2126 -6.8343 Source: Authors. Table B3: Change in Consumption, Business as Usual vs. Benchmark Category 2012 (%) 2018 (%) 2024 (%) 2030 (%) Food -1.4660 -2.1904 -3.9515 -6.6582 Household goods -1.6162 -2.7507 -4.8034 -7.7935 Consumer services -1.4708 -2.1687 -3.9112 -6.6060 Autos -1.5952 -2.4636 -4.3413 -7.1902 Electricity and LPG -1.6282 -2.7335 -4.7229 -7.6955 Public transport -1.5571 -2.5669 -4.5506 -7.4624 Gasoline -1.8873 -3.8617 -6.5138 -10.0627 Water -1.3804 -2.0513 -3.7879 -6.4516 Housing -1.4796 -2.0887 -3.8143 -6.4951 Source: Authors. 47 Table B4: Change in Imports, Business as Usual vs. Benchmark Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture -0.4702% -1.5789% -2.4229% -3.2258% Livestock 0.0000% -3.5714% -2.9412% -2.4390% Forestry 0.0000% 0.0000% 0.0000% -5.7143% Fisheries -0.0744% -0.2339% 0.0000% 0.0000% Oil 0.0523% 0.1704% 0.2401% 0.2951% Natural gas 0.0486% 0.1980% 0.2822% 0.3036% Mining 0.0000% -1.2048% -2.0202% -2.5316% Refining 0.0321% 0.1271% 0.1829% 0.2296% Transport 0.0000% -0.5405% -1.3575% -1.5038% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics 0.0094% 0.0323% 0.0468% 0.0590% Services -0.5270% -1.8847% -2.7855% -3.6547% Manufacturing -0.2625% -0.9855% -1.5178% -2.0147% Source: Authors. Table B5: Change in Exports, Business as Usual vs. Benchmark Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 2.4259% 1.7978% 0.3788% -1.9324% Livestock 1.8182% 1.5152% 0.0000% -2.1739% Forestry 0.0000% 0.0000% 0.0000% 0.0000% Fisheries 0.0000% 0.0000% 0.0000% -4.1667% Oil -0.0340% -0.2003% -0.3416% -0.4922% Natural gas -0.0297% -0.2415% -0.4197% -0.5169% Mining 2.1277% 1.7857% 0.0000% -2.5641% Refining -0.0160% -0.1400% -0.2500% -0.3678% Transport 2.2251% 1.0989% -0.7435% -3.5714% Electricity 0.0000% 0.0000% -6.6667% -5.5556% Chemicals & plastics 0.0111% -0.0288% -0.0698% -0.1182% Services 2.4612% 2.2068% 0.7964% -1.5336% Manufacturing 2.2743% 1.3939% -0.4409% -3.1834% Source: Authors. C. International comparison of subsidies across selected countries Table C1: Fuel subsidies (US cents) per kilowatt-hour (kWh), 2008 Data Country Oil product (US cents/kWh) Petroleum Producing Countries * 3.35 Argentina 2.51 Mexico 2.47 Colombia 0.71 China 0.67 Note: Petroleum producing countries include Algeria, Angola, Azerbaijan, Ecuador, Iran, Iraq, Kuwait, Libya, Nigeria, Qatar, Russian Federation, Saudi Arabia, Turkmenistan, UAE, Uzbekistan, and Venezuela. The lower bound figure for the group is 1.26 US cents/kWh and the upper bound figure is 6.45 US cents/kWh. Source: Charles and Wooders (2011). Table C2: Fossil fuel consumption subsidy rates as proportion of the full cost of supply Country Average subsidization rate, 2010 48 Mexico 12.5 Argentina 22 Colombia 4.3 China 3.8 India 13.5 Source: IEA, World Energy Outlook, 2011a. D. Trade under Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual Table D1: Change in Imports, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 0.0000% -0.6519% -1.1970% -1.7257% Livestock 0.0000% 0.0000% -2.9412% -2.4390% Forestry 0.0000% 0.0000% 0.0000% -5.7143% Fisheries 1.2638% 2.0939% 0.0000% 0.0000% Oil 0.0664% 0.1979% 0.2699% 0.3218% Natural gas -0.0150% 0.1082% 0.2123% 0.2552% Mining -0.7246% -1.8182% -2.0202% -2.5316% Refining 0.0163% 0.1025% 0.1662% 0.2207% Transport 0.6452% 0.0000% 0.0000% -0.3717% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics 0.0122% 0.0385% 0.0536% 0.0652% Services -0.5270% -1.8847% -2.6902% -3.4135% Manufacturing -0.3001% -1.1032% -1.6927% -2.2251% Source: Authors. Table D2: Change in Exports, Non-Revenue-Neutral Gradual Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 1.3624% 0.6818% -0.1905% -1.1182% Livestock 0.0000% 0.0000% 0.0000% -1.0753% Forestry 0.0000% 0.0000% 0.0000% 0.0000% Fisheries 0.0000% -5.8824% -5.0000% -4.1667% Oil -0.0587% -0.2471% -0.3883% -0.5181% Natural gas 0.0265% -0.1213% -0.2863% -0.4806% Mining 2.1277% 1.7857% 0.0000% 0.0000% Refining -0.0053% -0.1122% -0.2168% -0.3187% Transport 0.9284% -0.3344% -1.4032% -2.6751% Electricity 0.0000% 0.0000% -6.6667% -5.5556% Chemicals & plastics -0.0009% -0.0397% -0.0713% -0.1011% Services 1.6794% 1.8354% 1.2682% 0.4203% Manufacturing 1.4727% 1.0773% 0.3422% -0.6976% Source: Authors. E. Partial Neutrality Table E1: Change in Aggregate Results, Partial Neutrality of Subsidy Phaseout vs. Business as Usual 49 2012 2018 2024 2030 Category (%) (%) (%) (%) GDP -0.9830 0.1581 0.4775 1.1608 Investment -2.6421 -3.9031 -0.6494 16.9448 Government 5.3269 5.3153 4.5038 3.4095 Capital Stock — — — 7.1530 Aggregate welfare — — — 0.2151 Agent 1 — — — 3.3922 Agent 2 — — — 1.3617 Agent 3 — — — 0.1232 Agent 4 — — — -0.6228 Source: Authors. Note: — = results not reported. Table E2: Change in Production, Partial Neutrality of Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture -1.7590 -1.3802 1.1449 9.0006 Livestock -1.4162 -0.5734 2.2426 10.1736 Forestry -0.8658 0.3597 2.9851 10.4651 Fisheries -1.3333 -2.2727 -1.4354 2.3622 Oil -5.2448 -8.4355 -6.8040 0.0603 Natural gas -5.2356 -8.3945 -6.7895 0.0842 Mining -3.4682 -2.5641 1.4113 12.7644 Refining -8.9149 -15.0253 -12.6628 -3.9585 Transport -1.0222 -1.0830 0.2132 3.8149 Electricity -24.3648 -35.8347 -33.3097 -25.7069 Chemicals & plastics -5.2377 -6.1929 -1.3393 13.8088 Services -0.3105 0.0072 1.2251 4.6656 Manufacturing -1.6589 -0.2496 3.6028 13.9359 Source: Authors. Table E3: Change in Consumption, Partial Neutrality of Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Food -0.0390 -0.1166 0.5285 1.7232 Household goods -0.3297 -0.4267 0.2644 1.5076 Consumer services -0.5873 -0.6348 0.0078 1.2103 Autos -0.8941 -0.9589 -0.2516 0.9958 Electricity and LPG -8.2232 -12.8535 -12.1622 -10.8071 Public transport -0.3975 -0.7132 -0.0750 1.1062 Gasoline -1.9236 -3.0713 -2.4106 -1.2337 Water -0.1536 -0.1284 0.5382 1.7210 Housing -0.5644 -0.6003 0.0000 1.2107 Source: Authors. Table E4: Change in Imports, Partial Neutrality of Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 0.0000% -0.7833% -1.3072% -1.9108% 50 Livestock 0.0000% 0.0000% -2.9412% -2.4390% Forestry 0.0000% 0.0000% 0.0000% -5.7143% Fisheries 1.2506% 2.0282% 0.0000% 0.0000% Oil 0.0664% 0.1969% 0.2689% 0.3200% Natural gas -0.0163% 0.1056% 0.2105% 0.2546% Mining -0.7246% -1.8182% -2.0202% -2.5316% Refining 0.0163% 0.1009% 0.1651% 0.2199% Transport 0.6452% 0.0000% 0.0000% -0.3717% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics 0.0122% 0.0383% 0.0536% 0.0650% Services -0.6596% -1.9978% -2.6902% -3.4938% Manufacturing -0.2947% -1.0851% -1.6737% -2.1996% Source: Authors. Table E5: Change in Exports, Partial Neutrality of Subsidy Phaseout vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 1.6304% 0.6818% -0.1905% -1.4423% Livestock 1.8182% 1.5152% 0.0000% -1.0753% Forestry 0.0000% 0.0000% 0.0000% 0.0000% Fisheries 0.0000% -5.8824% -5.0000% -4.1667% Oil -0.0563% -0.2449% -0.3895% -0.5221% Natural gas 0.0299% -0.1175% -0.2859% -0.4873% Mining 2.1277% 1.7857% 0.0000% 0.0000% Refining -0.0053% -0.1122% -0.2168% -0.3240% Transport 1.1905% -0.1112% -1.4981% -3.0806% Electricity 0.0000% 0.0000% -6.6667% -5.5556% Chemicals & plastics 0.0019% -0.0397% -0.0736% -0.1070% Services 1.9414% 1.9595% 1.1114% -0.0222% Manufacturing 1.6987% 1.1131% 0.1172% -1.2643% Source: Authors. 51 F. Non-Neutral Elimination of Energy Subsidies and Expansion of VAT to Food and Medicines Table F1: Change in Aggregate Results, No Energy Subsidies and VAT vs. Business as Usual Category 2012 (%) 2018 (%) 2024 (%) 2030 (%) GDP -1.2090 -2.1677 -0.3520 1.1320 Investment -2.9719 -4.1835 -0.7584 17.3127 Government 7.2978 7.2883 6.4925 5.4213 Capital stock — — — 7.1195 Total welfare (∑Agent 1–4) — — — -0.4328 Agent 1 — — — 0.2957 Agent 2 — — — 0.2953 Agent 3 — — — -0.2487 Agent 4 — — — -0.9274 Government welfare — — — 4.3589 Aggregate welfare — — — 0.3960 Source: Authors. Note: — = results not reported. Table F2: Change in Consumption, No Energy Subsidies and VAT vs. Business as Usual Category 2012 (%) 2018 (%) 2024 (%) 2030 (%) Food -1.5050 -1.5812 -0.9295 0.2875 Household goods -1.0499 -1.1537 -0.4502 0.7967 Consumer services -1.1252 -1.1692 -0.5198 0.6873 Autos -1.2926 -1.3538 -0.6313 0.6168 Electricity and LPG -9.0237 -13.7306 -13.0174 -11.6215 Public transport -1.1479 -1.4594 -0.8118 0.3820 Gasoline -2.3940 -3.5547 -2.8810 -1.6997 Water -0.9288 -0.9056 -0.3257 0.9132 Housing -1.0830 -1.1207 -0.5063 0.6997 Source: Authors. Table F3: Change in Production, No Energy Subsidies and VAT vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture -2.7589 -2.3757 0.2363 8.4251 Livestock -2.2099 -1.3482 1.5272 9.7765 Forestry -1.3043 -0.3623 2.4024 10.2564 Fisheries -2.7027 -3.4483 -2.4155 1.1952 Oil -5.6670 -8.8442 -7.2009 -0.1208 Natural gas -5.6782 -8.8514 -7.2055 -0.0844 Mining -4.0698 -3.1941 0.9119 12.6984 Refining -9.4748 -15.6114 -13.2156 -4.2146 Transport -1.6954 -1.7509 -0.4291 3.2910 Electricity -24.9886 -36.4676 -33.9260 -26.0670 Chemicals & plastics -6.1764 -7.1894 -2.2379 13.4900 Services -0.4643 -0.1400 1.0977 4.6196 Manufacturing -2.2013 -0.7699 3.1793 13.8656 Source: Authors. 52 Table F4: Change in Imports, No Energy Subsidies and VAT vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 0.0000% -0.7833% -1.3072% -1.9108% Livestock 0.0000% 0.0000% -2.9412% -2.4390% Forestry 0.0000% 0.0000% 0.0000% -5.7143% Fisheries 1.2466% 2.0250% 0.0000% 0.0000% Oil 0.0669% 0.1974% 0.2694% 0.3203% Natural gas -0.0186% 0.1036% 0.2089% 0.2539% Mining -0.7246% -1.8182% -2.0202% -2.5316% Refining 0.0140% 0.1009% 0.1651% 0.2199% Transport 0.6452% 0.0000% 0.0000% -0.3717% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics 0.0124% 0.0385% 0.0536% 0.0650% Services -0.6596% -1.9978% -2.6902% -3.4938% Manufacturing Source: Authors. Table F5: Change in Exports, No Energy Subsidies and VAT vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 1.6304% 0.6818% -0.1905% -1.4423% Livestock 1.8182% 1.5152% 0.0000% -1.0753% Forestry 0.0000% 0.0000% 0.0000% 0.0000% Fisheries 0.0000% -5.8824% -5.0000% -4.1667% Oil -0.0568% -0.2460% -0.3900% -0.5226% Natural gas 0.0320% -0.1151% -0.2835% -0.4859% Mining 2.1277% 1.7857% 0.0000% 0.0000% Refining -0.0053% -0.1068% -0.2168% -0.3240% Transport 1.1905% -0.2227% -1.4981% -3.1621% Electricity 0.0000% 0.0000% -6.6667% -5.5556% Chemicals & plastics 0.0019% -0.0397% -0.0736% -0.1070% Services 1.9414% 1.9285% 1.1114% -0.0222% Manufacturing 1.6865% 1.1029% 0.1085% -1.2717% Source: Authors. G. Universal Social Insurance Proposal Mexico faces a highly segregated labor market, with a formal and a large informal sector. The informal sector does not have the same social insurance as the formal sector. Antón, Hernández, and Levy (2012) describe this in depth. They propose an innovative scheme under which the formal and the informal sector workers would have access to the same bundle of social insurance, deterring people and firms from joining the informal sector by lowering the costs of belonging to a more productive formal sector. They put forward a proposal whereby the current system is replaced by a complete removal of employee-employer contributions to social security. This would have the effect of moving labor from the informal sector (which does not contribute to social security) to the more efficient formal sector (where it is presently taxed to provide security and health benefits). To make up for the loss in revenue, Antón, Hernández, and Levy recommend including food and medical goods and services under the VAT (they are now not subject to such tax). To prevent any deleterious effect on equity and the 53 distribution of income, they provide a transfer to the lowest-income consumers to compensate for the higher cost of food and medical supplies that would be incurred. They employ a sophisticated theoretical model to demonstrate the economic soundness of their proposal, and then use an extensive dataset to quantify their theoretical results and test the robustness of their proposals. The model they employ, however, is static in nature. For this exercise, informality is defined as not contributing, on a sustained basis, to a formal social security system. The data we use for the structure of the labor market are shown in table D1. The costs involved in their proposal include health insurance (Mex$394.9 billion), retirement pensions (Mex$137 billion), and life and disability (Mex$28.1 billion), adding up to roughly Mex$560 billion. Table G1: Share of Informal Labor in each Sector, 2008 Informality Sector (% of labor force) 1. Agriculture 0.95 2. Livestock 0.95 3. Fishing 0.85 4. Forestry 0.95 5. Mining -0.06 6. Oil and gas 0.67 7. Refining output 0.78 8. Chemicals & plastics -0.04 9. Manufacturing 0.11 10. Electricity 0.01 11. Transport 0.07 12. Services 0.41 Sources: Authors. Table G2: Comparing USI and other expenditures, 2008 Concept Total amount (millions of pesos) Percentage of PIB PIB (2) 12,181,256 IMSS (1) 215,469 1.77 USI 559,300 4.60 Seguro Popular (3) 34,603 0.28 Programmable Expenditure (4) 1,899,925 15.60 Programmable Expenditure on Health 288,093 2.37 (4) Source: Antón et al 2012. H. Trade under Dual Policy Implementation vs. Business as Usual Table H1: Change in Imports, Dual Policy Implementation vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture 2.2866% 1.0256% 0.0000% -0.6289% Livestock 0.0000% 0.0000% -2.9412% -2.4390% Forestry 0.0000% 0.0000% 0.0000% -2.7778% Fisheries 1.5043% 2.2281% 0.0000% 0.0000% 54 Oil 0.0749% 0.2003% 0.2689% 0.3252% Natural gas 0.0366% 0.1434% 0.2326% 0.2561% Mining 0.7143% -1.2048% -2.0202% -2.1008% Refining 0.0387% 0.1196% 0.1752% 0.2248% Transport 1.2821% 1.0638% 0.4444% 0.0000% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics 0.0131% 0.0407% 0.0573% 0.0725% Services 0.1309% -1.5470% -2.5000% -3.4135% Manufacturing -0.5048% -1.3031% -1.8873% -2.5103% Source: Authors. Table H2: Change in Exports, Dual Policy Implementation vs. Business as Usual Sector 2012 (%) 2018 (%) 2024 (%) 2030 (%) Agriculture -2.8409% -0.4598% 0.7547% 0.9390% Livestock -1.8868% 0.0000% 1.2658% 2.0833% Forestry 0.0000% 0.0000% 0.0000% 6.6667% Fisheries -7.1429% -5.8824% 0.0000% 0.0000% Oil -0.0871% -0.2415% -0.3574% -0.4786% Natural gas -0.0433% -0.1587% -0.2929% -0.4372% Mining 0.0000% 1.7857% 1.4706% 2.4390% Refining -0.0497% -0.1232% -0.2009% -0.2879% Transport -1.6327% -0.1112% 0.3676% 0.2294% Electricity 0.0000% 0.0000% 0.0000% 0.0000% Chemicals & plastics -0.0191% -0.0346% -0.0531% -0.0756% Services -0.6643% 2.2068% 3.1865% 3.4941% Manufacturing -0.0189% 2.0211% 2.6159% 2.6658% Source: Authors. I. International Comparisons of CO2 Emissions per GDP Table I1. CO2 emissions from fuel combustion, 2009 (kg CO2 / US dollar (2000 prices)) CO2 emissions/GDP CO2 emissions/GDP using exchange rates using purchasing power parities China 2.17 0.55 India 1.81 0.35 Turkey 0.72 0.32 Chile 0.63 0.33 Mexico 0.55 0.36 United States 0.46 0.46 Colombia 0.43 0.14 Argentina 0.42 0.27 Brazil 0.39 0.20 Source: IEA 2011b. 55