«Will border carbon adjustments work? Niven Winchester*,†,‡, Sergey Paltsev* and John Reilly* Abstract The potential for greenhouse gas (GHG) ...»
Two historical cases are directly relevant for BCAs. First, in U.S.-Taxes on Petroleum and Certain Imported Substances (the Superfund case) a GATT dispute settlement panel allowed BTAs on chemicals contained in imported petroleum products. However, the panel did not specifically state that the substance had to be physically present in the final product (Neumayer, 2001). Second, in the late 1980s, the U.S. introduced a tax on ozone-depleting chemicals (ODCs) See, for example, Bhagwati and Mavroidis (2007), Bordoff (2009), Biermann and Brohm (2005), Frankel (2009), Goh (2004), Green and Epps (2008), Hoerner (1998), Brewer (2008), Pauwelyn (2007), Ismer and Neuhoff (2007) and WTO-UNEP (2009).
in order to implement the Montreal Protocol on Substances that Deplete the Ozone Layer. The tariff was applied to both ODCs and products containing or produced using these chemicals, but the legality of such measures is uncertain as the tariffs were never challenged under WTO rules (Brack et al., 2000).
GATT Article XX, which details general exceptions, provides another avenue to argue for BCAs. Two relevant exceptions include Article XX(b) and Article XX(g). Article XX(b) allows import restrictions that violate trade rules to be applied if they are necessary to protect human, animal or plant life or health, and Article XX(g) relates to the conservation of exhaustible natural resources.4 The process for determining the legality of GHG border measures is that, once implemented, countries “harmed” by the measures would need to lodge a complaint with the WTO, which would result in a ruling by the Dispute Settlement Body. In the absence of such a judgment, in remaining sections, we set aside legal issues and assume that BCAs are allowable under one or more of the above categories.
3. MODELING FRAMEWORKWe assess the economic and leakage impacts of BCAs using version four of the MIT EPPA model. EPPA is described in detailed by Paltsev et al. (2005) and we outline the core features of the model below. EPPA is a multi-regional, CGE model of the global economy that links GHG emissions to economic activity, and is solved through time in recursive dynamic fashion in fiveyear increments. There is a single representative utility maximizing agent in each region that derives income from factor payments and emissions permits and allocates expenditure across goods and investment. There is also a government sector in each region that collects revenue from taxes and purchases goods and services. Government deficits and surpluses are passed to consumers as lump sum transfers.
As illustrated in Table 1, EPPA recognizes Agriculture, five energy sectors (Coal, Crude oil, Refined oil, Gas and Electricity), two manufacturing sectors (Energy intensive industry and Other industry), Transportation and Services. Each good is produced by perfectly competitive firms that assemble primary factors and intermediate inputs. All goods are traded internationally and, following Armington (1969), goods are differentiated by region of origin using a constant elasticity of substitution function, except for Crude oil (which is treated as a homogenous commodity). Alternative electricity generation technologies in EPPA enhance abatement options.
Electricity can be produced using conventional technologies (e.g., electricity from coal and gas) and technologies not currently in use but which may become profitable as the emissions price rises (e.g., large scale wind generation and electricity from coal or gas with carbon capture and storage). As also indicated in Table 1, primary inputs include three non-energy resources and seven energy resources. Capital and labor are free to move between sectors and land is specific to agriculture. Each energy resource is sector specific. Crude and shale oil resources are perfect See Buck and Verheyen (2001) and Heinzerling (2007) for a discussion of legal issues associated with BCAs under Article XX.
substitutes in the oil sector, and the hydro, nuclear and wind & solar resources are specific to electricity generation technologies.
EPPA tracks the use of energy commodities (Coal, Refined oil and Gas) used in each sector measured in exajoules. These data combined with emissions per-exajoule coefficients for each energy commodity allow the model to predict (CO2) emissions. EPPA also traces non-CO2 GHGs (e.g., methane, and nitrous oxide) measured in CO2 equivalent (CO2-e) units using global warming potential (GWP) weights. GWP weights measure the ability of non-CO2 gases to trap heat in the atmosphere relative to the heat-trapping capability of CO2 over a 100 year period.
When GHG emissions are restricted, the model calculates a shadow value associated with the emission constraint, which is analogous to an emission price that would develop under a capand-trade program. The model is calibrated using economic data from the Global Trade Analysis Project (GTAP) database (Dimaranan, 2006) and energy balance data from the International Energy Agency (IEA).
3.1 Embodied GHGs and BCAs As noted above, H.R. 2454 does not set out how embodied GHG emissions will be calculated.
Following Rutherford and Babiker (1997), we use a comprehensive approach where total GHG emissions embodied in each commodity are the sum of direct and indirect emissions.5 Direct emissions are immediately linked with production, such as the combustion of fossil fuels to produce energy. Indirect emissions are associated with production of products used as intermediate inputs. For example, total emissions for automobiles equal emissions from the consumption of energy used in automobile manufacturing plus emissions associated with the
production of steel and other intermediate inputs. Our calculations employ equation (1):
X= AX + D (1) where X is an N×1 vector of total emissions per dollar for each of the N commodities; A is an N×N matrix, the ijth element of which is the number of dollars of good i used per dollar of good j being produced; and D is an N×1 vector of sectoral direct emissions coefficients per dollar of output.
Assuming that imported intermediate inputs embody the same quantities of GHG emissions as intermediate inputs sourced domestically, total embodied GHG emissions are computed by
solving (1) for X:
X= (I-X)-1D (2) BCAs are determined by embodied GHG emissions, calculated using equation (2), and carbon prices. For each applicable trade flow, we select an ad valorem tariff on imports of sector i from region s to r, τi,s,r, so as to increase the price of imports from s by the additional costs incurred by region s producers if they faced the carbon price in r. That is, τs is determined simultaneously
with the carbon price so that:
τi,s,rpi,s = pcarbrxi,s (3) where pi,s,r is the price of sector i in region s, pcarbr is the CO2-e price in region r, and xi,s is perdollar emissions embodied in production of i in s.
Embodied GHG calculations and BCA assignments are updated at the end of each modeling period but we do not adjust BCAs to account for the distribution of emission allowances specified in H.R. 2454, as such allowances are lump-sum transfers and will not influence firm behavior in our model.
3.2 BCA scenarios Yardsticks for BCA simulations are provided by business as usual (BAU) and cap-and-trade (CAT) scenarios, which we source from EPPA’s evaluation of the Energy Modeling Forum’s Climate Change Control Scenario described in Paltsev et al. (2009). In BAU, population and labor productivity advance at predetermined rates and there are no GHG restrictions, but autonomous improvements in energy efficiency and responses to rising energy prices as resources deplete lead to GHG emissions growing at a slower rate than GDP. The CAT scenario used in Paltsev et al. (2009) gradually reduces U.S. GHG emissions to 80% below 2000 We focus on the impact of BCAs that accurately target embodied GHG emissions and ignore monitoring costs. If monitoring costs are high, broad-brush trade measures may be preferred to targeted instruments (Engle, 2004).
emissions between 2015 and 2050, progressively reduces emissions in other Annex 1 regions except the Former Soviet Union (FSU) to 50% below 1990 levels between 2010 and 2050, and restricts emission in the FSU, China, India and Central and South America beginning in 2030. As it is unlikely that BCAs will be imposed after regions such as China and India begin to price emissions and EPPA has a five-year time step, we focus on the period prior to 2030 when only Annex 1 regions, excluding the FSU, implement climate policy. For ease of reference, we refer to this group as the “coalition” of nations implementing climate policy. In the period we analyze, emission allowances in the CAT scenario for each coalition region in each period match those in Paltsev et al. (2009). By 2025, relative to 2000, the U.S. reduces emissions by 31% and other coalition regions curtail emissions by between 18% and 35%. For simplicity, we do not allow banking of emission allowances over time.6 Although H.R. 2454 proposes BCAs on imports from regions where GHGs are not taxed with exemptions for some regions, we consider tariffs on imports from all non-coalition regions to simplify the analysis. Also, due to EPPA’s coarse sectoral aggregation, no sector in our model meets eligibility criteria for BCAs set out in H.R. 2454. However, the Bill’s BCA provisions are clearly aimed at manufactured products, or a subset of these commodities, so we consider BCAs for this sector, where manufacturing is defined as Energy intensive industry and Other industry.
To gauge the impact of the sectoral selectivity of BCAs in H.R. 2454, we also simulate tariffs on imports from non-coalition nations for all sectors. Leakage and competitive concerns also exist elsewhere, so in other simulations we consider BCAs imposed by all coalition regions, both on all sectors and manufacturing independently. That is, we analyze four BCA scenarios where, in addition to emission restrictions outlined in CAT, tariffs are imposed on imports from noncoalition regions: U.S. tariffs on all sectors (U.S.-ALL), coalition tariffs on all sectors (CLTALL), U.S. tariffs on manufacturing (U.S.-MNF), and coalition tariffs on manufacturing (CLTMNF).
4. MODELING RESULTS We focus on results for 2025 as BCAs are largest in this year. To understand what is driving our results, Table 2 presents total embodied GHG emissions by sector and region in 2025 calculated using equation (2). Emissions are reported in millions of metric tons (Mt) of CO2-e per U.S. billion dollars of output. Electricity produces significantly more GHG emissions per dollar of output than other sectors in most regions. Electricity GHG emissions per dollar are highest in China, where 34.4 Mt CO2-e are released per billion dollars of output. Emissions per dollar are also relatively high in Agriculture and Energy intensive industry. The numbers in Table 2 do not distinguish between gases, but unreported calculations reveal that agriculture emissions are largely non-CO2 gases while emissions from Energy intensive industry are predominantly CO2. Embodied emissions are also relatively high in Other industry, especially in non-coalition regions.
Note that these carbon constraints are more stringent than those in the H.R. 2454 because there are no credits from outside the capped sectors in the Energy Modeling Forum scenario.
Comparing carbon emissions across countries indicates that production in China and the FSU is relatively emission intensive and, in general, emissions per dollar are higher in non-coalition regions than coalition regions. However, care should be taken when making cross-country comparisons as the commodity composition of sectors may vary across regions and the number of physical units included in billion dollar bundles depends on the purchasing power of the U.S.
dollar relative to local currencies. Nevertheless, emissions per dollar coefficients in Table 2 are appropriate for calculating BCAs. For example, if agriculture production in a region is concentrated in GHG intensive commodities, exports from this region will produce more emissions than exports from a region that specializes in agriculture commodities that are less GHG intensive. Additionally, other factors constant, if a billion dollars buys twice as many units in region A as region B, one billion dollars of imports from A will embody twice the amount of emissions as imports from B.
Table 2. Embodied GHG emissions (CO2-e millions Mt per billion dollars), 2025.