This Selected Issues paper analyzes fiscal multipliers in Mexico. Estimates of fiscal multipliers--obtained from state-level spending--fall within 0.6-0.7 after accounting for dynamic effects. However, the size of multipliers varies with the output gap. The planned fiscal consolidation-under the estimated multipliers-is projected to subtract on average 0.5 percentage points from growth over 2015-20. However, there are offsetting effects. The positive growth impulse of lower costs on manufactured goods production is estimated to reach 0.5 percentage point in 2015 and 2016, largely offsetting the impact of fiscal consolidation on growth in the near term.

Abstract

This Selected Issues paper analyzes fiscal multipliers in Mexico. Estimates of fiscal multipliers--obtained from state-level spending--fall within 0.6-0.7 after accounting for dynamic effects. However, the size of multipliers varies with the output gap. The planned fiscal consolidation-under the estimated multipliers-is projected to subtract on average 0.5 percentage points from growth over 2015-20. However, there are offsetting effects. The positive growth impulse of lower costs on manufactured goods production is estimated to reach 0.5 percentage point in 2015 and 2016, largely offsetting the impact of fiscal consolidation on growth in the near term.

A Carbon Tax Proposal for Mexico1

Mexico has set clear commitments to reduce carbon emissions ahead of the 2015 Paris summit on climate change. A recent reform—pending Senate approval—proposed fixing excises on gasoline and diesel—in lieu of the current system of variable excises—and to allow the domestic price of these fuels to move more closely in line with international prices until they are fully liberalized in 2018. Taking into account climate change, air pollution, and traffic-related negative externalities associated with fossil fuel combustion, this paper finds that the proposed level of excises are close to optimal. However, natural gas and coal would remain significantly under taxed, if taxation is to become the main tool to correct for negative climate change externalities. Setting carbon taxes at optimal levels in Mexico could reduce carbon emissions by 6 percent over the medium term, contributing non-trivially to the goal of reducing carbon emissions in 25 percent by 2030.

A. Introduction

1. Fossil fuels are a critical energy input in industrial production and in final consumption, but their combustion has also harmful environmental and public health impacts. These consequences can have potentially sizable costs to the economy. However, energy prices in many countries are often set at levels that do not reflect these costs or are even subsidized. Given the seriousness of the environmental and public health problems associated with fossil fuel combustion, addressing them with targeted policy instruments is critical.

2. Mexico has been at the forefront among emerging markets in setting clear commitments to address these concerns. Mexico enacted in 2012 a climate change law which aimed at creating a legal framework to ensure the implementation of policies aiming at protecting the environment. It also created the Climate Change National Institute in charge of measuring carbon emissions. In 2013, Mexico released the national strategy for climate change, outlining clear commitments to reduce carbon emissions. In the same year, Mexico introduced a carbon tax on all fossil fuels as part of a broader fiscal reform. Recently, Mexico reiterated its commitments to reduce carbon emissions ahead of the Paris summit on climate change to take place in late 2015.

3. Fiscal instruments targeted directly at the sources of environmental harm help address these concerns but also generate fiscal revenues. As long as these revenues avoid increasing other taxes that distort economic activity, environmental protection is achieved at the lowest overall cost to the economy.

B. Mexico’s Current Excise Taxes on Fossil Fuels

4. Most fossil fuels in Mexico are subject to excise taxes, including a carbon component. A carbon tax refers to a tax directly linked to the level of carbon dioxide (CO2) emissions. The tax is often expressed per ton of CO2 and varies across fossil fuels depending on the amount of emissions they produce when burned. The carbon component of fuel excises in Mexico was introduced as part of the 2013 tax reform, as a special tax on production and services (IEPS) which came to effect in 2014. It covers all fossil fuels-except natural gas and crude oil (IEPS law, art. 8), and Jet fuel2—with rates varying depending on the carbon content. A second component of fuel excises in Mexico is earmarked for states and local governments (IEPS law, art. 2o-A II), and affects only gasoline (regular and premium) and diesel. A third component is a variable excise which also affects only gasoline and diesel. It is a residual between an international reference price converted to local currency and the regulated retail price.3 Since the international reference price co-moves one-for-one with international oil prices, the variable excise component moves countercyclically with the international price of oil, leading to higher fiscal revenues when international oil prices fall.

A04ufig1

Fuel Price Structure and Gasoline Prices

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: Secretary of Energy; and IMF staff calculations.

5. As of 2014, fossil fuel excise taxes in Mexico implied the lowest taxation rate among OECD countries. Implicit carbon taxes, defined as all excise taxes on fossil fuels, were low in Mexico, in particular as the variable excise mechanism implied a subsidy on fossil fuels until late 2014.

A04ufig2

Implicit Carbon Tax Rates

(U.S. dollars per ton of CO2)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: EIA; and OECD.Note: Mexico as of 2014, all others as of 2013.

6. The collapse of oil prices since late 2014 turned the variable excise tax positive. Domestic retail prices of gasoline and diesel continued to be adjusted upwards as oil prices declined, although at a much slower pace in 2015 than in the previous year. Therefore; Mexico no longer subsidizes diesel and gasoline given current domestic and international fuel prices. However, if the current variable excise mechanism were to be kept, the variable excise tax could turn into a subsidy again if international fuel prices rise.

C. In Search for a New Energy Taxation Mechanism

7. A recent reform—pending Senate approval—proposed a mechanism of excise taxes to ensure fuel subsidies are eliminated permanently and smooth the transition to fully liberalized domestic fuel prices. Mexico plans to liberalize domestic fuel prices by 2018, which implies a necessity to replace the current system of excises with a new mechanism. Ideally, this new mechanism should ensure a proper taxation of fossil fuels to reflect environmental and public health costs. The 2016 budget proposed a replacement of the variable component of the excise tax at levels near those observed this year with a fixed excise system. As of the writing of this paper, the proposal had been approved by the lower chamber of Congress. To smooth the transition toward fully liberalized domestic fuel prices, the Ministry of Finance would define a price band during this transition period (2016–2017) that would allow domestic fuel prices to fluctuate more in line with international prices.

8. A fixed excise tax on fossil fuels—equivalent to a carbon tax—is also a viable option to tackle the national climate change strategy goals, which have been reinforced with the recently announced commitments ahead of the 2015 Paris summit on climate change. Mexico has pledged to reduce carbon emissions in 30 percent by 2020, and 25 percent by 2030, relative to baseline projections. A carbon tax is administratively simple since Mexico has already a system in place. Since a carbon tax raises the price of each ton of carbon emitted into the atmosphere, it raises the cost of emissions. Higher costs induce a market response throughout the entire economy, creating incentives for fuel users to shift to alternative ways of production and ultimately lower carbon emissions. While Mexico’s contributions to carbon emissions are still small globally, they have been rising faster than the OECD average over the last decade.

A04ufig3

Carbon Dioxide Emissions

(Million metric tons)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Source: OECD.

9. From a fiscal perspective, fixing excise taxes on fuels at an optimal level could help stabilize tax revenues without relying on other, more distortionary taxation. Carbon taxes allow the extraction of a double dividend. Not only do they help correct an important negative externality but they also allow raising fiscal revenues without relying on taxes that tend to distort economic activity. If the variable excise system were maintained, lower tax revenues as international oil prices recover could lead to pressures to make up the shortfall by raising revenues through distortionary taxation. Setting excise taxes at optimal levels can also lead to lower public spending on health in the long-term if lower combustion of fossil fuel leads to a lower incidence of pollution-related diseases, keeping everything else equal.4

D. Estimating Optimal Carbon Tax Rates on Fossil Fuels for Mexico5

10. The methodology considers three main categories of externalities associated with fossil fuel combustion: climate change, air pollution, and those related to traffic. It relies on country-specific data wherever available and applicable to estimate the social cost of fossil fuel consumption in an attempt to estimate an optimal Pigouvian tax. While for some externalities variation in country-specific factors such as level of income, population density, and traffic congestion lead to different optimal taxes, this is not the case for climate change. For climate change the associated damage from carbon emissions is the same regardless of the fuel combustion process or where emissions are released. The methodology then assumes a uniform cost across countries, per unit of carbon emissions. More specifically, each externality is dealt with as follows:

  • Climate change is assumed to lead to a monetary cost of US$ 38 per metric ton of CO2. The assumed cost is within the range of estimates provided by the US Interagency Working Group on the Social Cost of Carbon (2013 update), obtained from simulations of a range of models, under different discount rate assumptions.6 These estimates, however, vary widely. As an illustration, the average estimated cost for 2015 ranges from US$12 per metric ton of CO2, if the discount factor is set to 5 percent, to US$38 and US$58 per metric ton of CO2 if the discount factor is lowered to 3 and 2.5 percent respectively. This assumption, together with data on carbon content of fossil fuels, allows deriving carbon charges for all fossil fuels that would reflect the cost of climate-change externalities.

  • To estimate the cost of air pollution, the approach uses empirical information on the link between exposure and incidence of pollution-related diseases. Local air pollution can lead to a wide array of harmful environmental consequences. However, to focus on those that are most severe, the analysis centers on premature human mortality. The methodology first determines how much polluted air is inhaled by exposed population. Relying on existing empirical evidence, it then assesses how this pollution exposure affects mortality risks, accounting for factors, such as the age and health of the population, which determine the vulnerability of the exposed population to pollution-related illness. The next step is the monetization of health effects. This monetary value is used to express the resulting damage in terms of unit of fuels, taking into account the carbon content of each fuel. 7

  • The cost associated with externalities derived from traffic congestion involves three components: congestion costs, accident costs, and road damage. Congestion costs stem from the value of travel time. This is obtained first from estimating the average added travel delay to other road users and then converted into a monetary cost related to local wage rates, using city-level and country-level data (Parry and others, 2014). Therefore, countries with high income level will face a greater cost through value of travel time per unit of time. Accident costs take into account country-specific road-fatality risks as well as other components of accident risk (property damage, medical costs, and nonfatal injuries). Road damage is estimated by attributing a portion of road maintenance expenditures to trucks, though these costs are modest in relative terms.

A04ufig4

Tax Rate Due to Climate Change

(U.S. dollars per GJ)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: Parry et al. (2014); and IMF staff calculations
A04ufig5

Estimated Pollution-related Fatalities

(Number of people per million ton of CO2 emitted)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: Parry et al. (2014); and IMF staff calculations.1/ Includes motor fuels and heating fuels.
A04ufig6

Components of Corrective Tax Rate due to Traffic-related Externalities

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Source: Parry et al. (2014).

11. Estimates of an optimal carbon tax for Mexico reach, quoted in MEX$ per GJ, 18 for natural gas, 77 for regular gasoline, 97 for diesel, and 36 for coal.8 Local air pollution and traffic-related are the dominant externalities for gasoline and diesel. Natural gas, being a much cleaner fossil fuel, requires only a small tax to correct for air pollution externalities from its combustion, but it still requires a significant tax to correct for climate change externalities as its combustion releases significant carbon dioxide into the atmosphere. Coal, on the other hand, requires a higher tax than natural gas on account of both, air pollution and climate change externalities.

A04ufig7

Estimated Optimal Carbon Taxes for Mexico

(Pesos per GJ, 2015)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: Parry et al. (2014); and IMF staff calculations.

12. Estimated tax rates are likely to be underestimated and should be interpreted as a lower bound. Models with richer micro-foundations, taking into account the interaction of carbon taxes with labor income taxes, can lead to higher carbon tax rates (Parry and Small, 2005). In the specific case of regular gasoline in Mexico, these interactions can lead to a carbon tax 35-percent higher than the one shown above (Anton-Sarabia and Hernandez-Trillo, 2014). Moreover, the methodology employed in this paper considers only premature human mortality when assessing the consequences of air pollution. However, there are other non-fatal consequences that can also have important economic costs. For instance, there is empirical evidence suggesting that increased pollution reduces labor supply (Hanna and Oliva, 2015); it lowers labor productivity (Graff and Nidel, 2012); and it affects children’s learning (Sanders, 2012).

13. Proposed excise tax levels are close to optimal. In an attempt to correct for some of the underestimation of optimal tax levels by the baseline methodology, the estimated taxes are scaled up by 35 percent—roughly the estimated contribution of interactions with labor income taxes discussed above. For simplicity, it is assumed that the scaling factor is the same for diesel and gasoline. The resulting “adjusted” Pigouvian tax is shown in the next figure together with the proposed taxes in the 2016 budget—in the case of gasoline and diesel—and those currently in place for coal and natural gas.9 The methodology employed here offers no specific tax for premium gasoline, but its higher energy content would suggest lower taxes than regular gasoline. Natural gas and coal are significantly under taxed—if their externalities were to be addressed entirely through taxation—but the proposed levels for gasoline and diesel are close to optimal. Note that the assumed adjustment to the estimated Pigouvian tax still has unaccounted factors, such as the impact of pollution on labor productivity and supply, which implies that the adjusted tax should still be seen as a lower bound.

A04ufig8

Estimated Optimal Carbon Taxes in Mexico

(Pesos per GJ, 2015)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Sources: Parry et al. (2014); and IMF staff calculations.

E. Impact of Imposing Carbon Tax Rates

14. Fixing excises at the baseline optimal levels would avoid a decline in fiscal revenues of about ½ percent of GDP in the medium term if the current system of excises were kept. In the short-run, fuel demand tends to be highly inelastic. Mexico-specific studies find estimates of price elasticities ranging from -0.1 to -0.2 (Galindo, 2005; Crotte and others, 2010; and references therein).10 Using the midpoint of this range we estimate the impact on fuel demand—relative to the baseline11—from applying the estimated optimal carbon tax rates to the main fossil fuels—gasoline, diesel, and natural gas represent about 80 percent of PEMEX sales. Taking into account the impact of higher energy prices on demand, fiscal revenues from fuel excises would stabilize at around 1.2 percent of GDP.12 In contrast, the current system of variable excises would imply lower revenues as oil prices recover.

A04ufig9

Fiscal Revenues from Excises

(In percent of GDP)

Citation: IMF Staff Country Reports 2015, 314; 10.5089/9781513593432.002.A004

Source: IMF staff calculations.

15. Concerns about short-run costs of higher energy prices need to be weighed against alternative ways to deliver on fiscal consolidation. The government is committed to lower the public sector borrowing requirements from 4.6 percent of GDP in 2014 to 2.5 percent of GDP by 2018. The required fiscal adjustment will likely imply a negative growth impulse in the short run, regardless of the tools used to deliver on this promise. Taxing energy properly does not get around this implication but it helps achieve this goal without having to rely on other, more distortionary taxation, such as income tax. Furthermore, it creates incentives to move toward cleaner and cheaper sources of energy with positive long-term implications on growth and health.

16. Over the long-run, the suggested taxation can contribute to a reduction of about 6 percent in carbon emissions, relative to our baseline. Mexican-specific studies find long-run price elasticities for fuel demand ranging from below -0.5 to numbers close to -1 (Galindo, 2005; Crotte and others, 2010 and references therein), depending on different econometric methods, geographic level, and time periods. Estimates at the national level fall closer to the -0.5 estimate than to -1. We use then -0.5—also used in Parry and others (2014)—in our calculations to derive long-term implications on carbon emissions. Under this approach, taxation alone can contribute non-trivially to the recent pledge of reducing carbon emissions in 25 percent by 2030.

F. Conclusions

17. The proposed reform to fix excise taxes on gasoline and diesel to eliminate fuel subsidies permanently is a welcome development. The proposed reform—already approved by the lower chamber of Congress but pending Senate approval—would replace the current system of variable excises with a more conventional fixed excise per liter of gasoline and diesel. The reform is an important step in the process of liberalization of domestic fuel prices and developing a true domestic energy market. It also eliminates permanently inefficient fuel subsidies, which were in place until 2014, and avoid a decline in fiscal revenues under the current system once oil prices recover.

18. From an environmental perspective, the proposed excise levels are close to an optimal Pigouvian tax, and could reduce carbon emissions by 6 percent. This reduction in carbon emissions would contribute non-trivially to Mexico’s reiterated commitments ahead of the Paris summit on climate change to take place in late 2015. Needless to say, the estimates are subject to substantial amount of uncertainty and will need to be updated frequently as technology and empirical evidence evolve.

References

  • Anton-Sarabia, Arturo and Fausto Hernandez-Trillo, 2014. “Optimal Gasoline Tax in Developing, Oil-producing Countries: The Case of Mexico.Energy Policy, vol. 67, pages 567571.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bovenberg, A. Lans and Goulder, Lawrence H., 1996. “Optimal Environmental Taxation in the Presence of Other Taxes: General-Equilibrium Analyses.American Economic Review, 1996, 86(4), pp. 9851000.

    • Search Google Scholar
    • Export Citation
  • Crotte, Amado, Robert Noland, and Daniel Graham, 2010. “An Analysis of Gasoline Demand Elasticities at the National and Local Levels in Mexico.Energy Policy, vol. 38, pages 44454456.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hanna, Rema and Paulina Oliva, 2015. “The Effect of Pollution on Labor Supply: Evidence from a Natural Experiment in Mexico City.Journal of Public Economics, vol. 122, pages 6879.

    • Search Google Scholar
    • Export Citation
  • Galindo, Luis, 2005. “Short- and long-run demand for energy in Mexico: a cointegration approach.Energy Policy, vol. 33, pages 11791185.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Graff, Joshua and Matthey Neidell, 2012. “The impact of pollution on worker productivity.American Economic Review, vol. 102, pages 36523673.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Parry, Ian, Dirk Heine, Eliza Lis, and Shanjun Li, 2014. “Getting Energy Prices Right.International Monetary Fund, Washington D.C.

    • Search Google Scholar
    • Export Citation
  • Parry, Ian and Kenneth A. Small, 2005. “Does Britain or the United States Have the Right Gasoline Tax? American Economic Review, vol. 95, No. 4, pages 12761289.

    • Search Google Scholar
    • Export Citation
  • Sanders, Nicholas, 2012. “What doesn’t kill you makes you weaker: prenatal pollution exposure and educational outcomes.Journal of human resources, vol. 47, pages 826851.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • US Interagency Working Group on the Social Cost of Carbon, 2013. “Thechnical Update of the Social Cost of Carbon for Regulatory Impact Analysis-Under Executive Order 12866.United States Government.

    • Search Google Scholar
    • Export Citation
1

Prepared by Fabian Valencia. The author thanks Dora Iakova, Luis Madrazo, Carlos Muñoz, Ian Parry, Glenn Sheriff, and seminar participants at the Mexican Ministry of Finance and the Central Bank of Mexico for insightful comments and suggestions and Alexander Herman for outstanding research assistance. This paper used Parry and others (2014) comprehensive calculations of optimal carbon taxes and updates it to 2015 prices.

2

Natural gas was exempted on account of being the cleanest fossil fuel. A 2013 Presidential Decree exempted Jet fuel from the carbon tax in adherence to international agreements signed by Mexico, which included the exemption of this fossil fuel from taxes.

3

The regulated retail price includes the carbon tax, excises earmarked for states, VAT, transport costs, distribution margins, and technical losses due to evaporation.

4

Parry and all (2014) calculate a potential reduction in deaths due to air pollution-related diseases of up to 20 percent for the case of Mexico.

5

This section briefly summarizes the methodology followed in Parry and all (2014) to compute Mexico’s suggested carbon tax rates. Further details and data sources can be found in Parry and all (2014) while theoretical underpinnings of fuel taxation can be found in Parry and Small (2005) and more general issues on optimal environmental taxation in the presence of other taxes in Bovenberg and Goulder (1996).

6

Quantifying climate change costs is subject to a great deal of uncertainty due in principal to three features of climate change. The first is the highly uncertain effect of emissions on specific climate outcomes; the second is the unknown form of human adaptation to problems building up over decades and centuries; and the third is the differences of opinion about the appropriate analytical procedure for aggregating effects occurring over long time intervals.

7

The population exposure uses an “intake fractions” approach, defined as the average pollution inhaled per unit of emissions released. This exposure is computed taking into account the population residing in regions within certain distance of emission sources; the change in the ambient concentration of pollution, for instance, per cubic meter; meteorological factors; and the emission rate of the pollutant in question (i.e. gasoline, diesel, coal, natural gas). Increased mortality rates from air pollution rely on empirical estimates on the link between intake fractions and mortality risks derived from pollution-related diseases. A controversial next step, but necessary to come up with a monetary cost of pollution, is to estimate how much people value mortality risk. Studies outside advanced economies are scarce. For this reason, the methodology takes an estimate of US$3 million from an OECD study (OECD, 2012) and adjusts it for differences in income levels to come up with country-specific values after updating the above benchmark to 2015 dollars. These monetary costs are ultimately expressed in terms of damage per ton of carbon emissions.

8

The estimates are computed in current US dollars, converted to MEX$ using purchasing power parity exchange rates, as recommended by Parry and others (2014). Purchasing power parity exchange rates take into account the local price level and thus reflect more accurately than market exchange rates people’s ability to purchase goods or pay out of their own income for risk reductions.

9

The 2016 budget proposed a reform to the IEPS law which encompasses substituting the variable excise component for gasoline and diesel with a fixed excise equal to the level shown in the graph. The proposed reform does not affect other fossil fuels, with natural gas continuing to be exempted.

10

Elasticities tend to vary across fuels, users, and even geographically; however, lacking enough granularity in projections of fuel consumption across users, regions, and fuel type, estimations are approximated using average elasticities.

11

The baseline projection for fuel demand is estimated using a value for the elasticity to real GDP growth of 0.3. Estimates of short-run income elasticities of fuel demand tend to be low. The chosen value of 0.3 in the construction of the baseline is within the range of values estimated in Crotte and others (2010).

12

For simplicity, we set the optimal tax of premium gasoline equal to the one on regular gasoline to compute estimated fiscal revenues.

Mexico: Selected Issues
Author: International Monetary Fund. Western Hemisphere Dept.