The Modular Applied GeNeral Equilibrium Tool (MAGNET) model, a multi-sector, multi-region computable general equilibrium model that covers the global economy, is used to evaluate several market-based mitigation policies to limit GHG emissions in agriculture. The policies analysed differ considerably in terms of the trade-offs they generate between mitigation outcomes and their associated impact on agricultural income, competitiveness, food consumption and government finances. This assessment provides policy makers with quantitative information about different policy design options that could deliver an acceptable blend of trade-offs, given their country-specific objectives and constraints.
Enhancing Climate Change Mitigation through Agriculture
2. Global analysis of mitigation policies for agriculture: Impacts and trade-offs
Abstract
The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.
The importance of agriculture to global mitigation efforts
As discussed in Chapter 1, agriculture contributes substantially to climate change and to mitigate global GHG emissions effectively and efficiently, agriculture must do its part. This will become increasingly important over time, given that agriculture has so far received less consideration in GHG mitigation policies compared with energy and other sectors (Bajželj et al., 2014[1]).
In the past, concerns about emissions leakage and loss of competiveness may have prevented countries from taking independent and early action. Such leakage occurs when mitigation policies in one region raise agricultural production costs and prices, causing supply from that region to fall, which creates incentives for increases in production and emissions elsewhere to partially fill the shortfall in supply.
In the absence of ambitious targets and policies for reducing agricultural emissions in most countries, this chapter explores how agriculture could make a substantial contribution to global mitigation efforts with a range of market-based policies. The potential economic consequences of policies to deliver ambitious emission reductions in agriculture, including their possible impact on competitiveness, food security and agricultural income, are also assessed.
A global tax on agricultural GHG emissions is the most ambitious policy option assessed, which assumes a willingness by all countries to apply an equally strong GHG tax rate, irrespective of their development status and priorities. This policy represents a high mitigation benchmark, which is then compared to a range of arguably more feasible but less effective policy options. These options include changing the burden of mitigation responsibility to exclude non-OECD countries, as well as applying the “beneficiary pays” principle rather than the “polluter pays” principle to incentivise ambitious mitigation outcomes for the agriculture sector. In recognition of the challenges and costs associated with measuring agricultural emissions, the efficacy of GHG-based payments on emission-intensive producer inputs and products is also examined.
With respect to evaluating the mitigation performance of different policy instruments, it is helpful to have in mind a reasonable or “fair” global emissions reduction target for agriculture globally. Taking into account relative mitigation costs and considerations about food security, Wollenberg et al. (2016[2]) suggest a non-CO2 emission reduction goal of 1 GtCO2eq yr-1 by 2030 for agriculture to contribute to the 2oC warming target by the end of the century. This represents an 11-18% reduction relative to the business-as-usual baselines assumed in their study and an allowable non-CO2 emissions budget of 6.15 to 7.78 GtCO2eq yr1. By comparison, a 1 GtCO2eq yr-1 emission reduction in this assessment represents a 14% reduction of the baseline emissions, bringing the baseline non-CO2 emissions in 2030 down from 7.33 to 6.33 GtCO2eq yr1. Wollenberg et al. (2016[2]) also propose a stronger longer-term target of and 2.5 GtCO2eq yr-1 by 2050 for agriculture’s contribution to meeting the 2oC target. These emission reduction targets have since become used as benchmarks in global mitigation assessments for agriculture, including in Frank et al. (2018[3]). Accordingly, they are used throughout this chapter as one of the performance benchmarks of the assessed policies.
In the next section, the model and data used for the analysis are described, along with the scope of the analysis and the selected mitigation policy instruments. Following this, the quantitative policy findings are presented. In the final section, the key policy messages and recommendations are explained, along with the main limitations of the analysis.
Modelling mitigation policies in agriculture for OECD countries and the world
The MAGNET model and scope of analysis
A computable general equilibrium (CGE) model is well suited to address many of the policy questions required to quantitatively assess the economic, competitiveness, and food security consequences of ambitious GHG mitigation targets for agriculture. A key strength of the CGE framework is its capacity to capture inter-sectoral relationships within agriculture, and between agriculture and other sectors, including other land use sectors. Other identified strengths include its ability to track trade relationships that influence competitiveness and leakage outcomes of mitigation policies, and the flow of costs and benefits to different sectors of the economy, including government, consumers and producers.
Given the utility of using a CGE model that can capture land use interactions with an acceptable degree of realism, the Modular Applied GeNeral Equilibrium Tool (MAGNET) model was selected (Woltjer and Kuiper, 2014[4]). This model has a long history of use within Wageningen University to assess the global impacts of policies in agriculture. It is a recursive dynamic multi-sector, multi-region Computable General Equilibrium (CGE) model that covers the global economy (Woltjer and Kuiper, 2014[4]). MAGNET is based on the Global Trade Analysis Project (GTAP) database and model developed by Purdue University in the United States (Hertel and Tsigas, 1997[5]). MAGNET and GTAP were originally designed to model the effects of trade policies, such as the Uruguay Round of multilateral trade negotiations, especially on the agricultural sectors. MAGNET has been extended and updated with several modules to improve the modelling of land markets and agricultural policies, biofuel policies, and the socio-economic and environmental impacts of environmental policies. There are eleven primary production sectors in agriculture, including eight crop sectors and three livestock sectors, and a total of 50 sectors in the model.
The version of MAGNET in this chapter uses the GTAP 9.2 database (Aguiar, Narayanan and McDougall, 2016[6]), which has a base year of 2011, but is updated in this assessment to create a dynamic baseline, from 2011 to 2050, with yield and economic growth assumptions that conform to the “middle of the road” Shared Socioeconomic Pathway (SSP2) (Fricko et al., 2017[7]). The model also incorporates emissions from the GTAP non-CO2 database (Irfanoglu and van der Mensbrugghe, 2015[8]), including methane (CH4) and nitrous oxide (N2O). This is complemented by CO2 emissions from the GTAP Energy-Environmental database (GTAP-E). Livestock non-CO2 emissions and Rice CH4 emissions are tied to the output variables of these respective sectors within the MAGNET model, whereas N2O emissions from crop fertiliser use are tied to the fertiliser input variable in these sectors. In addition, data on the marginal abatement costs (MACs) associated with practices and technologies that can be used to reduce GHG emissions are incorporated. These data are from the US Environmental Protection Agency (EPA) (2013) and they cover measures lowering the main non-CO2 emission sources, including methane (CH4) from enteric fermentation by ruminants (i.e. cattle, sheep and goats), nitrous oxide (N2O) and CH4 from livestock manure, CH4 emissions from paddy rice and N2O emissions from soil associated with fertiliser use by crops. Accordingly, it is these emission sources that are targeted by mitigation policies in this assessment. It should be noted that the MACs used in this assessment do not include assumptions about technological change from the development and adoption of new technologies which lower the costs of mitigation over time. Consequently, the MAC data used in this assessment are conservative with respect to their assumed GHG mitigation potential, especially over the longer term. The CO2 emissions associated with land use change (LUC) include changes in above and below ground carbon stocks between three aggregate types of land cover: cropland, grazing land, forest shrub land, and savannah land. The coefficients determining these changes in carbon stocks and CO2 emissions are drawn from the Agro-ecological Zone Emission Factor (AEZ-EF) model described in Plevin et al. (2014[9]).
In this assessment, GHG mitigation policies are only applied to non-CO2 emissions in the agriculture sector and not GHG emissions in other sectors of the economy. The possible implications of this modelling assumption are discussed below. Within agriculture, the vast majority of GHG emissions are targeted by most of the global mitigation policies considered in this assessment (78% of total agricultural GHG emissions in 2020, excluding LUC emissions). With reference to Figure 2.1, these include: CH4 from enteric fermentation and livestock manure management; N2O from livestock manure; N2O from fertiliser applied to crops; and CH4 from rice production. The remaining 22% of the emissions include CH4 and N2O emissions from the burning of biomass, and from fuel and energy use, and CO2 emissions from fuel and energy use. The emission sources that are not targeted by mitigation options considered here are still included in MAGNET and the changes in these emissions can be reported including, for example, changes in LUC emissions due to the expansion or contraction of agricultural land.
Figure 2.1. The agricultural GHGs in the MAGNET model (MtCO2eq), 2020
The economic impacts of mitigation policies on the different agricultural sectors and regions depend on the mitigation opportunities embedded in their MACs, and on the economic emission intensity of the sector’s output (i.e. the amount of GHG emissions from a sector divided by the economic value of its output). While there is a large variation in emission intensities across countries within a given sector, they are highest in the ruminant sector (OECD, 2019[10]). A GHG tax is therefore expected to have a relatively large impact on this sector.
Designing policies to unlock agriculture’s mitigation potential
Based on considerations about relevant and feasible mitigation policy options for agriculture, a set of eight mitigation policies was selected for assessment. These policies are considered sufficiently broad in scope to address the primary objective of identifying policy solutions that can unlock the large mitigation potential of the agricultural sector, without compromising food security in low-income regions while helping regions maintain their competitiveness. The first five policy options directly target agricultural emissions, whereas the last options target emission-intensive production inputs or consumer products.
The assessed policy instruments are listed below.
Policies that directly target emissions
Global tax on agricultural GHG emissions.
OECD tax on agricultural GHG emissions.
Global tax on agricultural GHG emissions combined with a food consumption subsidy
Global abatement payment for agricultural GHG emission reductions.
OECD abatement payment for agricultural GHG emission reductions.
Policies that target emission intensive production inputs and consumer products
Consumer-level GHG tax on ruminant meat and dairy products consumed within OECD countries.
Global GHG-based tax on emission intensive agricultural inputs, including ruminant animals and fertiliser
OECD GHG-based tax on emission intensive agricultural inputs, including ruminant animals and fertiliser
The first five policy scenarios listed above are assessed under dynamic settings, whereby the policies are applied from 2020 through to 2050. In each of these scenarios, the same increasing carbon price pathway is applied: with GHG prices of USD 40/tCO2eq, USD 60/tCO2eq, and USD 100/tCO2eq for the 2021-2030, 2031-2040, and 2041-2050 periods, respectively. These prices were considered to represent a reasonably high level of mitigation ambition compared to the much lower carbon market prices that have been observed to-date, where such markets exist. The USD 60/tCO2eq price approximately corresponds to the value that some modelling studies suggest will be required to limit temperature increases to 1.5°C (Rogelj et al., 2015[11]). For technical reasons related to the fact that the final three scenarios impose a GHG-based tax on consumer products or producer inputs, it was necessary to assess these scenarios in static mode.1 For these cases, 2050 was selected as the simulation year and a GHG price of USD 100/tCO2eq was applied in order to be consistent with the prices used in the other scenarios for this same year. The mitigation performance of the policies simulated under dynamic settings is evaluated with respect to their capacity to achieve the non-CO2 emission reduction targets of 1 GtCO2eq yr-1 by 2030, and 2.5 GtCO2eq yr-1 by 2050, proposed by Wollenberg et al. (2016[2]).
Beginning with the policies that directly target emissions, the first three follow the “polluter pays” principle by imposing a tax on emissions. The global taxes on GHG emissions, with and without the food consumption subsidy are the most ambitious policy options, as they assume a willingness by all countries to apply an equally strong GHG tax rate, irrespective of their development status and concerns about food production and food security. As mentioned above, the purpose of the first policy – the global tax on emissions – is to provide a high mitigation benchmark which can then be compared to a range of more feasible, but potentially less effective, mitigation policy options. In an attempt to address concerns that low-income countries may have about negative impacts on food production and agricultural incomes, a second scenario is defined where the tax on GHG emissions is limited to OECD countries. This option is, however, likely to erode the competitiveness of agriculture in OECD countries and cause a leakage of emissions mitigated by OECD countries into non-OECD countries. The third policy is a hybrid instrument that attempts to exploit the large mitigation potential that a global tax on agricultural emissions can provide by driving the restructuring of agricultural production in favour of sectors with lower GHG emissions, while at the same time providing a subsidy to consumers to maintain their baseline levels of food consumption.
The fourth and fifth policy options differ from the previous options by applying the “beneficiary pays” principle and providing an abatement payment to cover the mitigation costs of agricultural producers. This provides the same marginal abatement incentives as the GHG tax, but does not impose any tax burden on agricultural producers. The abatement payment is paid by the government to producers, and it precisely compensates producers for the costs they incur to reduce emissions at the selected carbon prices.
The final three scenarios are based on polices that attempt to circumvent the substantial challenge of measuring and monitoring GHG emissions from agricultural producers by applying a GHG-based tax to either emission-intensive production inputs (ruminant animals and fertiliser) or emission intensive consumer products (processed ruminant meat and dairy products). These policies would allow a saving in transaction costs (not quantified in this assessment) related to the measurement of emissions, but they would result in a loss of economic efficiency by failing to reward producers who lower their emissions by adopting mitigation practices that aim to lower emission intensities. The consumer-level GHG tax translates the value of emissions for the given tax rate into an equivalent tax set at the same rate for both domestic and imported consumer products within each OECD country or region, based on the economic emission intensity of the domestically-produced product. This tax is applied to ruminant meat and dairy products only. The motivation behind this policy is to address competitiveness and leakage issues that would typically emerge from the non-global application of a GHG tax by preserving the competitive position of domestic and imported products by taxing them at the same rate. This removes the onerous challenge of applying different tax rates to consumer products sourced from different destinations according to their emission intensities.
A notable omission from the above policy options is an emission-trading scheme. It is worth mentioning that an emission-trading scheme could be designed to provide similar mitigation and economic outcomes for agriculture as does the GHG tax and abatement payment mechanisms. According to economic theory, the auctioning of emission permits can provide the same mitigation incentives as a GHG tax, while the provision of free emission permits to agriculture could provide similar mitigation incentives as the abatement payment. Consequently, many of the insights on the mitigation effectiveness and economic impacts from the assessed instruments can be generalised to a broader range of market-based mitigation instruments than those assessed here.
GHG emission reductions and economic consequences of mitigation policies in agriculture
The quantitative impacts of the assessed policy instruments on emission reductions, agricultural producers and food consumers are presented. A more detailed regional breakdown of the modelling results is provided in the appendix of (OECD, 2019[10]).
The global GHG taxes, with and without the food subsidy, appear to be the most effective mitigation policies, narrowly missing the 1 GtCO2eq, non-CO2, 2030 mitigation target, and slightly exceeding the 2.5 GtCO2eq 2050 targets described in the previous sections (Figure 2.2 to Figure 2.4, Table 2.1). The global abatement payment is less effective, but still able to go about halfway towards achieving these targets. Although the GHG tax and abatement payments provide the same marginal mitigation incentives, the cost and price increases from the tax cause a contraction in the supply and demand for agricultural products in aggregate, but particularly from more emission-intensive sectors. This contraction is a major contributor to the overall reduction in emissions induced by this policy in some regions. For the ruminant sector aggregated across non-OECD countries, falls in production account for 42%, 43%, 46% of emission reductions of the global GHG tax in 2030, 2040 and 2050, respectively. Globally, the contribution of falling ruminant output to the total emission reductions of the ruminant sector is more muted at 28%, 26%, and 15%, respectively, as overall ruminant production in OECD countries increases over all three simulation periods. Accounting for the changes in LUC emissions reveals that the taxation policies could be substantially more effective by 2050 (Figure 2.3, Table 2.1). This results from a global shift in land cover from pasture to forest and shrub land, which will increase global carbon stocks over time as the ruminant grazing footprint contracts, particularly in Sub Saharan Africa and Latin America. Following the global abatement payment, LUC emissions increase relative to the baseline (Figure 2.4, Table 2.1), mainly due an increase in cropland at the expense of forest and shrub land in South East Asia and Latin America. However, these changes in land cover are one to two orders of magnitude smaller than the changes in land cover caused by the GHG tax. This nevertheless illustrates the potential importance of coupling this policy option with regulations to prevent the clearing of non-agricultural land containing comparatively high carbon stocks. Note that the consumer-level tax and tax on input policies are not displayed in Figure 2.3 and Figure 2.4 because they were only conducted for 2050.
As expected, the OECD GHG tax leads to the leakage of or increases in emissions in non-OECD countries, partially reducing its effectiveness (Table 2.1). The OECD GHG abatement payment is able to eliminate these leakage effects and provide a similar level of global mitigation as the OECD GHG tax, without the same negative consequences for agricultural production. Nevertheless, the policies confined to OECD countries make only small progress towards the proposed mitigation targets at the selected carbon prices (Figure 2.2 and Figure 2.3, Table 2.1).
The results of the consumer-level tax and tax on input policies that were assessed in static mode are presented in Table 2.2. For the purposes of comparison, the global GHG tax and the global abatement payment were also assessed in static mode for the year 2050 because dynamic and static scenario results cannot be meaningfully compared.2 The global tax on ruminants and fertilisers generated less than one-fifth of the emission reductions achieved by the global GHG tax and about two-fifths of the reductions from the global abatement payment. This is partly because the global tax on ruminants and fertilisers targets a smaller volume (86%) of the emissions than the global GHG tax and the abatement payment. When limited to OECD countries, its impact is naturally much smaller, with leakage effects further weakening its effectiveness.
Figure 2.2. Global reductions in agricultural non-CO2 emissions for dynamic policy scenarios
Figure 2.3. Global reductions in agricultural non-CO2 and land use change emissions for dynamic policy scenarios
Figure 2.4. Global reductions in agricultural non-CO2 and land use change emissions for the global abatement payment
Table 2.1. Summary of annual agricultural non-CO2 and LUC emission reductions policy instruments assessed under dynamic settings (MtCO2eq), in 2050
* The leakage rate is calculated as the sum of the increases in agricultural GHG emissions in non-OECD countries, divided by the sum of the reductions in agricultural GHG emissions in OECD countries.
The percentages of the baseline non-CO2 emissions reduced in each broad region are provided in parentheses.
The OECD consumer-level tax can negate the leakage of emissions, but as with the OECD ruminant and fertiliser tax, it is one of the least effective instruments for lowering emissions. The ineffectiveness of these less targeted approaches appears to worsen when the tax is levied at the consumer rather than at the input stage. This is because the impact of the tax is further weakened by the diversion of affected farm commodities from domestic to export markets, and by the diluting effect of intermediate inputs in the final processed food products.
The global GHG tax, abatement payment, and GHG tax with food subsidy, each have differing impacts not only on emission levels, but also on agricultural producers and food consumers. While the GHG tax leads to the largest emission reductions, it has the most detrimental effect on farm income (measured as value-added or returns to the land, capital and labour endowments, at agents prices), particularly in non-OECD regions. It also causes the largest reduction in food consumption (weighted by value at constant 2020 world prices), though not nearly as large as its impact on producers (Table 2.3). Conversely, it generates the largest increases in government revenue (Table 2.4).
Table 2.2. Summary of annual agricultural non-CO2 emission reductions for policy instruments assessed under static settings (MtCO2eq), 2050
Table 2.3 Changes in agricultural value-added and household food consumption from policies, 2050
Note: OECD regions are indicated in bold. North America consists of the United States and Canada. Europe covers all OECD European countries. Other OECD includes Japan, Korea, Israel, and Turkey. MENA-Caspian includes the Middle East, North Africa and countries of the Caspian region. East and South East Asia include China, South East Asia, and non-OECD countries in East Asia. Latin America includes all non-OECD Latin American countries.
Table 2.4. Annual changes to government budget from selected global GHG mitigation policies, 2050 (USD million)
A different but somewhat improved assembly of trade-offs emerges from the addition of a food consumption subsidy to the GHG tax. The combined policies have similar impacts on reducing emissions and on producers, but this time consumption is maintained and raises a smaller but still positive amount of government revenue in all regions apart from one. However, given the substantial negative impact of this policy on producers in low-income countries, it would be very likely to reduce food security for the rural poor in these same countries. Note that in Sub-Saharan Africa, the global GHG tax does not cause aggregate food consumption to fall. In this region, the crop sector benefits from the reduction in input prices that ensue from the substantial fall in emission intensive livestock production, expanding its production (OECD, 2019[10]). On balance, this has a positive net impact on aggregate, value-weighted, food consumption in 2050. Consequently, in this year, this region does not receive a food consumption subsidy in the GHG tax with food subsidy scenario. In all other simulation periods, aggregate food consumption weighted by value declines in all regions.3
The global abatement payment offers the prospect of appreciable global emission reductions (Table 2.1) without harming agricultural producers or food consumption at the aggregate regional level (Table 2.3). However, in contrast to the GHG tax policies, the abatement payment needs to be paid for. In this assessment, the cost of the abatement payment is paid by governments within each region. These policies not only differ in terms of who incurs the cost of abatement, but also with respect to the size of these costs, with costs of the abatement payment to government being much smaller than the cost of the GHG tax to producers. This asymmetry occurs because the abatement payment covers only the cost of reducing emissions, whereas the GHG tax is levied on the entire stream of producers’ non-CO2 emissions (i.e. both the abated and unabated portion of emissions).
In addition to their impact of food consumption, producer income, and government budgets, these instruments generate different economic welfare impacts. To assess these impacts, the welfare measure known as equivalent variation (EV) was used. This approach uses government expenditures as a proxy for welfare obtained from public goods (Keller, 1980[12]). It is also often used in CGE analyses to approximate changes in the efficiency with which economic resources are allocated within the economy. Global EV for the global GHG tax and the global abatement payment is USD -27 944 million and USD -18 430 million, respectively, in 2050. These figures are negative, indicating there is a loss of welfare associated with these policies that attempt to mitigate GHG emissions (World Bank, 2018[13]). The welfare loss from the tax is about 50% larger than the loss associated with the abatement payment; however, the tax generates about 100% and 300% higher non-CO2 reductions and total emission reductions (non-CO2 + LUC emissions), respectively (Table 2.4). Therefore, from an economic welfare perspective, the abatement payment performed worse than the GHG tax relative to the quantity of emissions reduced. However, this is a partial evaluation of the economic welfare because it does not consider welfare benefits in terms of the avoided damage costs associated with emission reductions achieved by each policy. If these benefits were considered, both policies could deliver an improvement in net welfare.
Another more policy targeted option, but which is not assessed in this chapter, would be to redirect part of the existing producer support provided to the sector for non-environmental purposes to pay for the abatement payment instrument. This approach to lower the sector’s carbon footprint is presently gathering support among international experts and agencies, including the World Bank (2018[13]). With 2015-2017 agricultural support for the 51 countries considered in the OECD’s Agricultural Policy Monitoring and Evaluation 2018 (2018[14]) calculated to be USD 484 billion, there are arguably sufficient resources to easily cover annual abatement payments for OECD and non-OECD countries, which are projected to reach USD 2 312 and USD 9 022 million, respectively, by 2030, and USD 5 675 and USD 25 117 million, respectively, by 2050.4 The financial burden of this instrument would increase further if a more ambitious carbon price path capable of reaching the sector’s 2030 and 2050 mitigation targets was assumed.
Other funding arrangements may be feasible, for example the purchasing of agricultural emission reduction credits by other sectors that are required to pay for emitting GHGs, notwithstanding the political challenges that may be associated with initiating such transfers. This approach would be possible in the few locations with operational emission trading schemes (e.g. the European Union and New Zealand, although more countries are expected to adopt national carbon pricing schemes in future).
To provide some validation of the model results it is useful to compare the magnitudes of emission reductions from this assessment with comparable global studies. The non-CO2 emission reduction potentials of 0.43-0.84 GtCO2eq at USD 40/tCO2eq in 2030, 0.81-1.57 GtCO2eq at USD 60/tCO2eq in 2040, and 1.33-2.71 GtCO2eq at USD 100/tCO2eq in 2050, from the global GHG tax and abatement payment policies assessed in this chapter, are well within the range of potentials from existing studies in the literature. According to the most recent Assessment Report of the Intergovernmental Panel on Climate Change (Smith et al., 2014[15]), annual emission reductions for agriculture of 0.03-2.6 GtCO2eq – at USD 50/tCO2eq,and 0.2-4.6 GtCO2eq at USD 100/tCO2eq in 2030 – are based on results from different studies (Rose et al., 2012[16]; McKinsey & Company, 2009[17]; Golub et al., 2009[18]; Smith et al., 2007[19]) and include soil carbon sequestration as well as non-CO2 emission reductions. A more recent partial equilibrium assessment by Frank et al. (2018[3])(2018) calculated higher non-CO2 mitigation potential in 2030 of 1 GtCO2eq at only USD 25/tCO2eq, but with a slightly lower mitigation potential of 2.6 GtCO2eq at USD 100/tCO2eq in 2050. These figures are comparable to those in this assessment, although the models differ significantly in structure and emission baselines and in the way they integrate abatement options.
There may also be substantial additional mitigation from changing consumers’ dietary choices to include a less emission intensive basket of food commodities (Bajželj et al., 2014[1]; Wollenberg et al., 2016[2]; Poore and Nemecek, 2018[20]). However, no clear or effective policy options have been proposed to achieve this. The hybrid policy assessed in this chapter, which combined a GHG tax with a food consumption subsidy, provides one option for incentivising such a dietary shift without sacrificing total food consumption. The assessment of this policy could, however, be improved by focusing on maintaining the nutritional value of consumption rather than its value at constant world prices.
As with all modelling assessments, there are caveats. For instance, the mitigation potential of the policies calculated in this chapter may be lower than the agricultural sector’s full potential because the mitigation policies only target 78% of the sector’s non-CO2 emissions. Moreover, the mitigation potentials for these emissions that are included are also conservative, because the MACs used do not consider technological changes that lower the costs of mitigation over time. In addition, options to sequester soil carbon in grasslands and croplands were not considered. This omission was due to the absence of reliable global data on the marginal costs of soil carbon sequestration.
Including mitigation policies in non-agricultural sectors, particularly land use sectors, could have important implications for the performance of mitigation policies in agriculture. Competition for land between agriculture and forestry can be particularly influential for agricultural production and emissions. Research by Golub et al. (2009[18]), also using CGE model, showed that subsidising carbon sequestration in the forestry sector can increase forest area at the expense of grazing land, causing extensive ruminant production and emissions to contract. When combined with a GHG tax on agricultural emissions, this contraction intensified. Considering mitigation more broadly for the land use sector as whole would be a useful extension to the assessment in this chapter.
Another caveat is the absence of climate change impacts in the baseline and policy scenarios. However, the policy insights from the assessment, in terms of the relative magnitudes of the different policies and the types of trade-offs they induce, are unlikely to change very much if these impacts were taken into account. At the global level, most studies assessing climate change impacts over time do not predict very large changes in agricultural production between now and 2050. For instance, Nelson et al. (2013[21]) project a mean global decline in crop production of only 2% by 2050, and van Meijl et al. (2018[22]) simulate a similar small decline in agricultural production of between 0.5 and 2.5% by 2050. Still, there will be larger impacts in some regions. Importantly, however, Meijl et al (2018[22]) found, in a model inter-comparison study covering five global models (IMAGE, CAPRI, GLOBIOM, MAgPIE, MAGNET) that non-CO2 emission taxes and land-based mitigation policies in agriculture, commensurate with the sector’s contribution to a 2oC global warming target, would have a much larger negative impact on agricultural production than the effects of climate change.
It would have also been instructive to assess the impact of transferring a portion of existing coupled support payments to agriculture to fund the GHG abatement payment. However, given that the level of support among countries is so variable, some countries could easily fund abatement this way, while others could not. Consequently, this approach could result in differentiated impacts, with countries that are able to transfer coupled support to abatement activities possibly experiencing stronger reductions in emissions and output as a consequence of removing support. Further work on quantifying these impacts is recommended, including the calculation of possible emission leakage effects that may arise from the ensuing adjustments in competiveness.
Summary of findings
There is growing recognition of the importance of reducing GHG emissions from agriculture to meet the ambitious targets of the Paris Agreement goal to limit global average temperatures to well below 2°C and pursue efforts to limit the increase to 1.5°C above pre-industrial levels. The challenge for policy makers is to find ways to reduce agricultural emissions in a way that also minimises the negative consequences of mitigation policies on food security, agricultural income, and competitiveness.
The policies assessed in this chapter differed considerably in terms of the trade-offs they generated between mitigation outcomes and their associated impacts on agricultural income, competitiveness, food consumption, and government finances. The mitigation effectiveness is assessed with reference to annual non-CO2 emission reduction targets of 1 and 2.5 GtCO2eq by 2030 and 2050. These are not official targets, but have been proposed by some analysts as being commensurate with agriculture’s global emission contribution and capacity to mitigate.
The global GHG tax-induced large emission reductions are more or less aligned with the above targets, but imposed the largest economic costs on agricultural producers, particularly in the emission intensive ruminant sectors of many developing countries. They also slightly reduce household food consumption, although it should be possible to insulate consumers from the associated negative impact linked to the resulting higher food prices by combining the tax with a food subsidy, which could be financed by the GHG tax. The global abatement payment offers the prospect of appreciable global emission reductions without harming agricultural producers or food consumers, although only half as effective as the tax in reducing non-CO2 emissions. The effectiveness of the abatement payment could fall further if emissions from land use change are also taken into account due to the small expansion of agricultural land that can result from this policy. From an economic welfare (or efficiency of economic resource allocation) perspective, the abatement payment performed worse than the GHG tax, relative to the quantity of emissions reduced.
Moreover, unlike the GHG tax policies, which generate government revenue, the global abatement payment would need to be funded. However, the level of payment needed globally represents a small proportion of the agricultural producer support currently provided by countries for non-environmental purposes.
The policy options which levy GHG taxes on emission proxies, such as more easily measurable and emission intensive production inputs or consumer products, were found to be far less effective than directly taxing emissions. Their ineffectiveness appears to worsen when the tax is levied at the consumer stage compared to the input stage.
The geographical scale of policies is critical to their mitigation effectiveness. More than a third of the GHG emission reductions from a GHG tax that is limited to OECD countries could be leaked as increases in emissions in non-OECD countries. If an abatement payment to OECD country emissions were applied instead, these leakage impacts could be controlled while delivering a similar level of global mitigation. However, it is clear that OECD countries alone cannot make a meaningful contribution to lowering global agricultural emissions given the dominant share of non-OECD countries in global agricultural emissions.
References
[6] Aguiar, A., B. Narayanan and R. McDougall (2016), “An Overview of the GTAP 9 Data Base”, Journal of Global Economic Analysis, Vol. 1/1, pp. 181-208, http://dx.doi.org/10.21642/jgea.010103af.
[1] Bajželj, B. et al. (2014), “Importance of food-demand management for climate mitigation”, Nature Climate Change, Vol. 4/10, pp. 924-929, http://dx.doi.org/10.1038/nclimate2353.
[15] Edenhofer, O. et al. (eds.) (2014), Agriculture, Forestry and Other Land Use (AFOLU, Cambridge University Press.
[3] Frank, S. et al. (2018), “Structural change as a key component for agricultural non-CO2 mitigation efforts”, Nature Communications, Vol. 9/1060.
[7] Fricko, O. et al. (2017), “The marker quantification of the Shared Socioeconomic Pathway 2: A middle-of-the-road scenario for the 21st century”, Global Environmental Change, Vol. 42, pp. 251-267, http://dx.doi.org/10.1016/j.gloenvcha.2016.06.004.
[23] Gernaat, D. et al. (2015), “Understanding the contribution of non-carbon dioxide gases in deep mitigation scenarios”, Global Environmental Change, Vol. 33, pp. 142-153, http://dx.doi.org/10.1016/j.gloenvcha.2015.04.010.
[18] Golub, A. et al. (2009), “The opportunity cost of land use and the global potential for greenhouse gas mitigation in agriculture and forestry”, Resource and Energy Economics, Vol. 31/4, pp. 299-319, http://dx.doi.org/10.1016/j.reseneeco.2009.04.007.
[5] Hertel, T. and M. Tsigas (1997), Structure of GTAP, Cambridge University Press, https://pdfs.semanticscholar.org/beeb/0c756facf0372b68ac215a6ee1160780c566.pdf.
[8] Irfanoglu, Z. and D. van der Mensbrugghe (2015), ““Development of the version 9 non-CO2 GHG emissions database”, Documentation accompanying dataset”, Center for Global Trade Analysis, Purdue University.
[12] Keller, W. (1980), Tax incidence: A general equilibrium approach, Amsterdam.
[17] McKinsey & Company (2009), Pathways to a low-carbon economy: version 2 of the Global Greenhouse Gas Abatement Cost Curve, McKinsey & Company, London.
[21] Nelson, G. et al. (2013), “Assessing uncertainty along the climate-crop-economy modeling chain”, PNAS, Vol. 111, pp. 3274-3279.
[10] OECD (2019), A Global Economic Evaluation of GHG Mitigation Policies for Agriculture, COM/TAD/CA/ENV/EPOC(2018)7/FINAL, OECD, Paris.
[14] OECD (2018), Agricultural Policy Monitoring and Evaluation 2018, OECD Publishing, Paris, https://dx.doi.org/10.1787/agr_pol-2018-en.
[9] Plevin, R. et al. (2014), “Agro-ecological Zone Emission Factor (AEZ-EF) Model (v47)”, GTAP Technical Paper, Vol. 34.
[20] Poore, J. and T. Nemecek (2018), “Reducing food’s environmental impacts through producers and consumers”, Science, Vol. 360/6392, pp. 987-992, http://dx.doi.org/10.1126/science.aaq0216.
[11] Rogelj, J. et al. (2015), “Energy system transformations for limiting end-of-century warming to below 1.5 °C.”, Nature Climate Change, Vol. 5, pp. 519-527, http://dx.doi.org/10.1038/nclimate2572.
[16] Rose, S. et al. (2012), “Land-based mitigation in climate stabilization”, Energy Economics, Vol. 34/1, pp. 365-380, http://dx.doi.org/10.1016/j.eneco.2011.06.004.
[24] Smith, P., et al. (2008), “Greenhouse gas mitigation in agriculture”, Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 363.
[19] Smith, P. et al. (2007), “Greenhouse gas mitigation in agriculture”, Philosophical Transactions of the Royal Society B: Biological Sciences, Vol. 363/1492, pp. 789-813, http://dx.doi.org/10.1098/rstb.2007.2184.
[22] van Meijl, H. (2018), “Comparing impacts of climate change and mitigation on global agriculture by 2050”, Environmental Research Letters, Vol. 13.
[2] Wollenberg, E. et al. (2016), “Reducing emissions from agriculture to meet the 2 °C target”, Global Change Biology, Vol. 22/12, pp. 3859-3864, http://dx.doi.org/10.1111/gcb.13340.
[4] Woltjer, G. and M. Kuiper (2014), The MAGNET Model, http://www.wageningenUR.nl/en/lei.
[13] World Bank (2018), Realigning Agricultural Support to Promote Climate-Smart Agriculture, World Bank Other Operational Studies 30934, World Bank.
Notes
← 1. For policies vi, vii and viii, small policy shocks are used to restrict the tax revenue generated by taxing either consumption or inputs to match the revenue that would be collected GHG tax on the emissions that are associated with these inputs and outputs. Given the small policy shocks, interaction with dynamic features is expected to be limited, so these shocks were implemented in static mode for the year 2050.
← 2. The reason is that the agriculture sector is exposed to mitigation incentives over a sustained period (2020-2050) in the dynamic scenarios, causing emissions to diverge quite considerably with the dynamic baseline. In contrast, the emission reductions achieved with a specific mitigation policy applied for single year to the 2050 baseline, as is done with the static simulations, are smaller than the emission reductions achieved in 2050 under dynamic settings.
← 3. Latin America experiences a similar pattern of production effects. However, the substitution effect between crops and livestock is not as strong and the share of the food that is derived from crop-based sources is lower than in Sub Saharan Africa. For this region, the global GHG tax causes aggregate food consumption to decline.
← 4. Note these 2050 figures for the non-OECD countries do not equal those presented in the table because the Russian Federation and non-OECD European countries are not included in the latter.