In this paper, we analyze and compare abatement costs associated to greenhouse gas emissions from agriculture in twelve EU countries. We also examine the ...
Agriculture and Climate Change in the European Union: Greenhouse Gas Emissions and Abatement Costs May 2001(*) Prepared for the AAEA Annual Meeting – Chicago, August 4-8 2001 Stéphane De Cara Fapri/Card Iowa State University, Ames, IA
Pierre-Alain Jayet UMR d'Economie Publique INRA ESR, Grignon, France Short abstract
In this paper, we analyze and compare abatement costs associated to greenhouse gas emissions from agriculture in twelve EU countries. We also examine the possibility offered to farmers to afforest CAP set-aside land and discuss the differences in the countries' interests to promote carbon sequestration in international negotiations. Abstract This paper addresses the assessment of greenhouse gas emissions from agriculture in the European Union. We first estimate and compare net emissions from agricultural activities in twelve EU countries. These estimates are based on a set of farm-unit linear-programming models. We then use these models to derive marginal and total abatement costs associated with different levels of reduction targets (dual approach) and different values of carbon-equivalent emissions (primal approach). Finally, we explore the possibility of allowing afforestation on setaside land. This paper highlights the discrepancies between countries regarding abatement costs and their sensitiveness to the accounting for carbon sequestration. JEL classification: Q25, Q28.
(*)
Copyright 2001 by Stéphane De cara and Pierre-Alain Jayet. All rights reserved.Readers may make verbatim copies of this document for non-commercial purposes by any means, provided that this copyright notice appears on all such copies
1
Introduction
Agricultural activities play a significant role in increasing the concentrations of greenhouse gases (GHG) in the atmosphere. As reported by the UNFCCC Greenhouse Gas Inventory Database, agriculture contributes for about 10 percent to European total emissions (year 1998, IPCCC, 2000). The two major gases emitted by the agricultural sector are nitrous oxide (N2O) and methane (CH4). Agricultural activities are responsible for approximately 50 and 60 percent of the European emissions in methane and nitrous oxide respectively. Moreover, UNFCCC figures do not account for carbon sequestration due to agricultural activities and changes in land use. In fact, by removing carbon dioxide (CO2) from the atmosphere, agricultural activities may contribute to slow down the accumulation process of GHG in the atmosphere and, consequently, reduce the threat on global climate. The costs to achieve a given reduction in GHG emissions are sensitive to the definition and the method used to account for carbon sequestration in national GHG inventories. Yet no agreement has been reached on this particular point in the international debate on climate change (IISD, 2000). By signing the Kyoto Protocol (1997), the EU countries committed themselves to reduce their emissions by 8 percent in 2008-2012 as compared to 1990 levels. They set up a burden-sharing rule, whereby each Member State is assigned a specific national target. As far as the EU is concerned, GHG emissions from agriculture have an importance that exceeds their relative weight in total emissions. Firstly, the agricultural sector, by contrast with that of energy for instance, remains among the economic sectors where the EU experience is the strongest as regards of regulation, and international negotiations. Secondly, such regulation policies may contribute to the greening of the CAP if they allow to achieving successfully supply regulation goals while complying with environmental objectives. Thirdly, they may benefit from existing administrative bodies related to the CAP, which can help to lower implementation costs. 1
The literature on climate change mitigation policies mainly focuses on energy-related emissions of CO2. However, recent studies show that multi-gas approaches could significantly contribute to lower the abatement costs by widening the portfolio of potential abatements (Manne and Richels, 2000; Hayhoe et al., 1999; Reilly et al., 1999; Burniaux, 2000). Due to the variety of sinks and GHG emissions that it is responsible for, agriculture is among the sectors for which the multi-gas approach appears to be the most relevant. Some studies provide estimates of abatement costs related to carbon sequestration, either in trees by using different land-use and forestry options (Plantinga et al., 1999; Newell and Stavins, 2000) or in soils through different tillage practices (Babcock and Pautsch, 2000). Nevertheless, in the perspective of a GHG regulation policy, one should take into account the multi-gas nature of GHG emissions from agriculture. To our knowledge, only a few studies integrate the various possible GHG sources and carbon sinks in agriculture (Simons et al., 1994 for Dutch agriculture; Schneider, 2000, in the case of the United States). In this paper, we extend a study by De Cara and Jayet (2000) for French agriculture to the EU scale. The objective of this paper is twofold. We first analyze emissions and abatement costs both on the EU scale and on a country-by-country basis. To do so, we use a set of farm-unit linear programming models, in which emissions are linked to producing activity levels. Abatement costs are derived for different rates of emission reductions (referred hereafter as the dual approach) and various levels of carbon value (primal approach). Second, we examine the possibility offered to farmers to afforest set-aside land. Along with fitting into EU supply regulation policy, it may provide farmers with another means to reduce their net emissions and lower their abatement costs. By underlining the discrepancies among countries’ interest to promote carbon sequestration, our results shed some light on the recent stalemate experienced at November 2000 UNFCCC Conference in The Hague on carbon sequestration issues. 2
2 2.1
The model The generic farm-type LP model
The generic model is based on linear programming (LP) methods (including integer and binary variables). Each model describes annual supply choices for a given farm type. The farm-type representation allows to accounting for the wide diversity of technical constraints faced by European farmers. The net emission levels for each source/sink are directly linked to the levels of activity, endogenously chosen by farmers. The primary source of data is the sample of the European Farm Account data Network (FADN). This sample is representative of about 2.5 millions of European farmers (full-time farming) in twelve European countries (France, Great Britain, Germany, Italy, Spain, Ireland, The Netherlands, Denmark, Greece, Belgium, Portugal, and Luxemburg). This sample has been divided into homogenous farm-types with respect to climatic conditions, soil characteristics and technical production possibilities. Thus, each farm-type belongs to a specific European region and corresponds to a given main producing activity. To reflect specific thresholds set up by the CAP measures, other criteria are used in the typology, such as crop yields, area allocated to various crops, and altitude. By this way, 472 farm-types (groups) are obtained, each being associated with a specific LP model. Each farm-type is viewed as a single firm representative of the whole group behavior. A producer of type k is supposed to choose his supply level and input demand (xk) in order to maximize his gross margin (πk) subject to production constraints (Ak⋅xk ≤ zk). Let P1k, the optimization problem for the k-th producer, be:
max π k ( x k ;θ k , φ ) ≡ k x
g ⋅ xk ≡ k
f ∈F
( p f − c kf ) x kf −
k k k k k k P1 s.t. A (θ ,φ ) ⋅ x ≤ z (θ ,φ )
xk ≥ 0
3
h∈H
c hk x hk +
i∈I
pi x ik −
j∈J
c kj x kj
A ∈ ℜ mxn
(C1)
x ∈ ℜn
(C2)
This problem is linear with respect to xk, the primal n×1-vector of the n activities. The sets F and H stand for crop activities. The first set denotes crops bound to be sold and the second one represents those that may be on-farm consumed (pastures, forages and feed grains). Sixteen crop producing activities are allowed in the model and represent most of the European agricultural land use, including activities for setting aside the different types of land as per CAP measures. I is the set of livestock activities, and the set J includes the set of purchased livestock feed grains. The m×n-matrix Ak and the m×1-vector zk contain respectively the input-output coefficients and the capacities of the m constraints on production. The 1×n-vector gk contains the crop margins. The vector of parameters θk characterizes the k-th type of producer whereas φ stands for the vector of general economic parameters not dependent on type k. The optimum levels of various variables are assigned an asterisk. The constraints can be divided into five types: (i) crop rotation; (ii) nutritional needs of cattle in terms of energy and proteins; (iii) initial endowments of quasi-fixed factors (land and livestock); (iv) bovine livestock demography; (v) restrictions imposed by the CAP measures. The composite and “either/or” nature of the options offered by the CAP set-aside policy is modeled through the use of integer and binary variables. The 3×1-vector E1k(x*k) stands for the emission levels of each greenhouse “gas” (methane, nitrous oxide and carbon sequestration) associated with activities in the optimal solution. The computation of the components of E1k is done after optimization on the basis of the relationships between activities and emissions described hereafter. The mapping function f(.) computes net carbon budget from the components of E1k. It integrates conversion coefficients into CO2equivalent for methane and nitrous oxide on the basis of 100 years GWP. Furthermore, CO2 results are converted into their carbon content by the ratio of CO2 weight to carbon weight. For the optimal solution, the net budget e1k in terms of C-CO2 is such that e1k = f(E1k(x*k)). 4
2.2
Evaluation of emissions 2.2.1 Sources of methane
Methane (CH4) emissions from agriculture are due to enteric fermentation. Animal feeding is a key-factor in the determination of the methane flow. Animal feeding is endogenously chosen by each representative farmer, this choice relying on the relative cost of feedstuffs and their nutritional characteristics. If feedstuffs are on-farm produced, the cost considered is an “opportunity”' cost. This choice is restricted by the necessity to respect minimum nutritional supply and by enteric capacities of animals. Following Sauvant et al. (1996), methane emissions can be computed by using the following two equations for simple feed grains and compound feedstuffs, respectively: E-CH4/EB = -1.73 + 13.91.dE
(1)
E-CH4/EB = 5.62 + 4.54.dE
(2)
where dE (in percent) stands for the digestibility and E-CH4/EB (in percent) stands for the share of gross energy food value loss in methane. The protein, energy and digestibility characteristics of feedstuffs have been taken from Jarrige (1988). 2.2.2 Sources of nitrous oxide Following Bouwman (1989), nitrous oxide (N2O) emissions are linked to amounts of nitrogen (N, in kilograms) brought. The per-hectare and per-year flow of N2O (in kilograms) is computed as follows: N2O = 1.88 + 0.00417. N
(3)
2.2.3 Carbon sequestration Three types of crops are distinguished according to their carbon storage in soils: main grains (0.4 tC.ha-1), pastures (0.6 tC.ha-1) and forests (0.75 tC.ha-1). These figures are taken from 5
Balesdent's study (1995), assuming fallow land as the starting point. As for forests, an additional 2.5 tC.ha-1 is assumed for afforested land in addition to carbon storage in soils. Thus, afforested land is assumed to store an average annual flow of 3.25 tC. ha-1 (soils and trees). This pertains to the average annual increase over a complete rotation. In the model, farmers choose among the different forestry activities available according to the discounted revenues they yield (see De Cara and Jayet, 2000). 2.3
Evaluation of marginal abatement costs 2.3.1 Dual approach
In order to evaluate the marginal abatement cost for a given type of producer, the problem as formulated in P1k is solved. It provides a baseline estimation of initial emissions, e1k. Then, P1k is transformed into Pαk by adding a new constraint (C3): f(E1k(xk)) ≤ α.e1k
λk(α)
(C3)
where α < 1. The shadow-price (taken at the optimum) associated to constraint (C3) is denoted by λ *k(α). It gives the marginal loss of gross margin, for the k-th producer and for a specified percentage of (1-α) cut-off of net emissions. In other words, the marginal abatement cost for a given value α..e1k of emissions is represented by the implicit cost faced by the k-th farmer implied by constraint (C3). The total abatement cost for a given α is πk(x*k(1)) - πk(x*k(α)), where x*k(1) and x*k(α) stand for the non-constrained and the new primal solution vector, respectively. By gradually decreasing α, we obtain the level and the slope of the total abatement cost functions for each representative subgroup k and each set of α.e1k quantities of abatement. This procedure allows us to explore the differences among producers' abatement costs for various given relative decrease of European agricultural emissions. Note that α is a uniform percentage of emission cut-off across all the farmers. Thus, unless initial 6
emission levels are the same for all k, the burden should be far different from one producer to another for the same α. Indeed, the same rate of reduction corresponds to different quantities of reduction in emissions. This method directly provides in the same time individual and European rates of reduction in emissions. Yet, it does not directly show how the concentration of emissions among farmers for a specific rate of reduction would be influenced by a regulation policy. In order to give a good representation of individual cost functions in terms of abatement quantities, the α-range has to be wide enough and the difference between two consecutive steps, small enough. This method should not be seen as an environmental “command-and-control” policy. It should be seen rather as a method allowing the environmental agency to know how marginal abatement costs are distributed around an assumed value of social marginal damage. The lower is α, the higher should be λ*k(α) and the loss in πk(x) since it becomes gradually more expensive –both marginally and absolutely– to respect the constraint imposed by the inequality (C3). Consequently, the convexity of abatement costs implies that the cumulative function of marginal abatement cost per farm should move to the right as the percentage of reduction increases. In other words, for α1 > α2, the percentage of firms for which we have λ*k(α2) ≤ λ (λ given) should be lower than the percentage of firms whose λ*k(α1) is such that λ*k(α2) ≤ λ. It is obvious that, in a first-best world and for a given social value of carbon, the percentage of abatement will vary across farmers. In the optimum, marginal abatement costs should be equal across individuals and there is no reason that optimal decisions correspond to the same rate of reduction in initial emissions. Thus, the analysis of the cumulative functions shown in figure 1 should consider all the spectrum of possible reduction rates and not only one single cumulative function associated to a given rate of reduction of initial emissions. 7
2.3.2 Primal approach In this section, we describe briefly a more commonly used approach to estimate the abatement costs. We introduce a tax t on individual net emissions and reformulate each linear program as follows:
k
max π k ( x k ; θ k , φ ) ≡ g ⋅ x k − t.e( x k ) ) k
P
k t
x
s.t.
A k (θ k , φ ) ⋅ x k xk
≤ ≥
z k (θ k , φ ) A ∈ ℜ mxn (C1) 0 x ∈ ℜ n (C2)
The net emissions in nitrous oxide and methane and the carbon sequestration in soils are based on the relationships exposed above. At the optimum, the marginal abatement cost equates the level of the tax. With t varying, we thus obtain marginal abatement cost functions for each farmtype. Unlike the dual approach, this alternative approach puts the emphasis on the relationship from prices to quantities. It allows a more direct analysis of impacts of a regulation policy, but necessitates an a priori assumption of carbon value.
3 3.1
Results Evaluation of emissions
We first derive the initial levels of emissions from the models P1k and aggregate the results for each European country. This first set of results calls for some caveats and cautionary comments, as they may differ slightly with IPCC estimates (see figure 1). First, the relationships used in the computation of emissions are not exactly the same as those used in the UNFCCC inventory database. In the case of methane for instance, we focus on the link with animal feeding, rather than solely on animal numbers. Second, as for nitrous oxide emissions, only mineral fertilizer use is considered in the model (organic fertilizer use is neglected because of the lack of reliable data). Third, our results are affected by the setting of the FADN sample. Actually, the FADN 8
sample is known to be rather unrepresentative for some categories of farmers (part-time farming) and its comprehensiveness varies from one country to another. Therefore, total emissions of methane and nitrous oxide emissions are slightly underestimated compared to IPCC estimates. However, our results give a good representation of the sharing of agricultural emissions within the EU. As reported in the IPCC inventory, we find that French emissions total a quarter of European emissions in CH4 and N2O. Likewise, the four largest emitters (France, Great Britain, Germany, and Italy) account together for more than two third in European emissions.
80
70
60
MtCO2
50
40
30 20 10 0 FR
GB
GE
N2O (IPCC)
IT
SP
IR
ND
CH4 (IPCC)
DK
GR
N2O
BE
PT
LU
CH4
Figure 1: Comparison of the results of the model with IPCC estimates for methane and nitrous oxide emissions (IPCC estimates from www.ghg.unfccc.int: enteric fermentation (CH4) and agricultural soils (N2O), year 1994).
9
3.2
Dual approach 3.2.1 European Union
We compute the empirical cumulative curves of marginal abatement costs as exposed above. We examine the marginal abatement costs for rates of reduction from 1 percent to 20 percent of the initial levels of individual net emissions. These curves are shown in figure 2 for the entire European Union. For a low 1 percent rate of reduction, 90 percent of the farmers face a marginal abatement costs lower than EUR 200. That means that, if a carbon tax of EUR 200 were implemented (in a first-best world with no uncertainty and no monitoring costs), 90 percent of the farmers would be able to achieve at least a 1 percent reduction in their net emissions. For a carbon value of EUR 400 per ton, all the farmers would be able to reduce their emissions by at least 1 percent. For the same rate of reduction, one half of the farmers face a marginal abatement cost lower than EUR 48. Symmetrically, for the highest rate of reduction examined (20 percent), the median corresponds to a carbon value of EUR 325. The marginal value of carbon should be at least EUR 600 in order that 90 percent of the farmers reduce their emissions by 20 percent. The reverse interpretation of figure 2 emphases the relationship from prices to quantities. For instance, for an arbitrary marginal value of carbon of EUR 70, the farmers’ population can be divided as follows: about one third of the farmers would be able to reduce their emissions by less than 1 percent; one half of the farmers would be able to reduce their emissions by between 1 and 10 percent, the remaining sixth being able to reduce their emissions by more than 15 percent.
10
1 1% 5% 10% 15% 20%
Percentage in total population
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
100
200 300 400 500 600 700 Marginal abatement cost (euros)
800
900
Figure 2: Cumulative functions of per-farm marginal abatement costs.
Obviously, the total potential of emission reduction is not straightforward with this method, as it depends on the distribution of initial emissions. Therefore, the next step in this analysis consists in identifying some relevant characteristics of the farmers located on each cumulative curve associated to a specific rate of reduction. For each rate of reduction, we sort the farmers according to their marginal abatement costs and divide the total population into ten classes of identical weight. Note that individual abatement cost functions and requested quantities of abatement differ from one farmer to another. Thus, the composition of the classes may also differ from one rate of reduction to another. Within each class and for four rates of reduction, we compute the average initial emissions and 11
the average marginal abatement costs (weighted by the numbers of farmers within each group). The results are presented in table 1. Consistently to the intuition, the farmers for which the abatement costs are the lowest for low rates of reduction (first class in table 1) are also those who are responsible for the highest emissions. For these farmers the potential of reduction at low cost is generally larger because of the large quantities of emissions in the initial situation. Symmetrically, as for farmers belonging to the tenth class (largest marginal abatement costs), emission levels are relatively low, which indicates a scarcity of low-cost potentials of reduction. Nevertheless, the negative relationship between initial emissions and marginal abatement costs becomes less and less apparent, as the requested rate of reduction increases. 1% Class 1 2 3 4 5 6 7 8 9 10
λ 2.4 9.8 19.7 31.3 39.9 53.6 75.9 98.7 143.5 262.7
5% e1 147.5 105.5 90.5 106.5 76.2 106.9 88.6 64.4 68.3 64.3
λ 8.9 26.8 45.3 62.1 95.3 129.1 163.2 197.9 269.8 497.8
10% e1 130.1 94.3 62.1 76.1 95.5 98.7 95.0 129.7 84.1 52.8
λ 16.3 51.8 77.3 119.0 154.8 204.7 265.2 323.9 405.3 652.5
15% e1 129.1 63.0 81.9 72.8 98.0 111.6 109.3 90.3 86.2 70.1
λ 24.9 61.0 111.1 165.0 220.1 281.0 330.7 390.5 470.6 716.1
20% e1 122.5 52.4 86.7 95.5 107.4 90.2 101.3 71.6 118.6 69.4
λ 36.2 92.5 177.0 242.5 305.0 352.2 388.3 446.5 533.4 813.7
e1 103.2 69.5 114.1 96.9 89.6 110.8 50.3 106.3 101.1 76.5
74.5 91.1 150.8 91.1 227.5 91.1 280.4 91.1 341.0 91.1 Average Table 1: Average initial emissions and marginal abatement costs for different rates of reduction in emissions.
3.2.2 Country comparison Table 2 compares the marginal abatement costs and the emissions on a country-by-country basis. For a given rate of reduction, the lowest marginal abatement costs can be found in Portugal, Ireland and France. The average marginal abatement cost in these countries is lower than the 12
European average, suggesting that these countries should contribute the most to the European effort to reduce GHG emissions from agriculture. By contrast, the highest marginal abatement costs occur in the Netherlands for rates of reduction ranging from 5 to 20 percent. This is mainly due to the importance and intensiveness of livestock producing activities in this country. Number of farms
Emissions
Marginal abatement cost λ15%
λ20%
(EUR) 216.8 316.3 311.2 350.3 267.6 196.3 546.3 280.0 224.0 462.3 187.9 285.1
(EUR) 272.2 399.4 354.8 403.2 354.1 258.3 671.1 347.6 295.2 579.7 208.8 300.3
Total 2,461.2 224.3 91.1 74.5 150.8 227.5 280.4 Table 2: Country comparison of initial emissions and marginal abatement costs.
341.0
Country France Great Britain Germany Italy Spain Ireland Holland Denmark Greece Belgium Portugal Luxemburg
3.3
(,000) 362.7 128.4 278.2 549.5 313.4 131.1 69.8 56.8 272.6 39.4 257.7 1.7
Total
per-farm
(Mt CO2) 56.8 34.5 34.2 28.0 18.0 14.6 11.0 5.6 6.8 5.9 8.6 0.4
(tCO2) 156.5 268.4 122.9 51.0 57.3 111.6 157.7 97.7 25.1 149.7 33.2 239.9
λ1%
λ5%
λ10%
(EUR) (EUR) (EUR) 69.5 114.6 171.7 48.0 161.8 261.1 59.4 106.7 222.6 97.3 204.2 288.6 83.8 177.9 219.6 67.7 91.7 183.4 86.1 325.6 445.0 74.8 178.3 238.4 81.4 133.7 208.3 65.6 197.6 317.5 45.3 84.0 140.9 135.3 217.9 391.8
Primal approach
The marginal abatement costs are now computed by using an explicit value of carbon emissions in each LP program. For each farm type, net emissions are thus viewed as a new costly activity that enters the objective function. The cost associated with a unit of net emissions is supposed to be t, which we allow to vary in the range [0, EUR 400] per ton of C-CO2 (by steps of EUR 5). 3.3.1 Influence on the different sources of GHG emissions Figure 3 shows the influence of a first-best tax on net emissions of methane and nitrous oxide, 13
and a subsidy to carbon sequestration. The influence of a tax on net emissions is the strongest in the case of methane. This signals lower abatement costs for reduction in emissions in this gas. The most striking feature of figure 3 (left) is the high sensitivity of methane emissions to a tax ranging from EUR 10 to EUR 50. In this zone, substitutions in animal feeding allow reduction in emissions at low-cost. Beyond this point, substitutions in animal feeding are not sufficient and farmers have to reduce animal numbers, which raises abatement costs. Our results also highlight the relative rigidity of nitrous oxide emissions. The latter result is consistent with those found by Schneider (2000) for the US. As for carbon storage, the variations remain in a narrow range (less than 0.5 MtCO2). Two effects actually affect the evolution of carbon storage. As the first-best tax increases, forest and pasture activities become increasingly attractive. Nevertheless, the relatively low differential in carbon storage for the various producing activities tends to restrict the substitution from low to high-carbon potential activities. Conversely, reductions in methane emissions are obtained through changes in animal feeding, and mainly through a move from on-farm consumption to purchased animal feeding. These substitutions tend to decrease the area allocated to pastures and forage, and thus lower carbon sequestration. Europe
Europe
135
-29.55 N20 CH4
130
StC -29.6 -29.65
Emissions (MtCO2)
Emissions (MtCO2)
125 120 115 110 105
-29.7 -29.75 -29.8 -29.85 -29.9
100
-29.95
95
-30 0
50
100
150 200 250 Tax (euro/tCO2)
300
350
400
0
50
100
150 200 250 Tax (euros/tCO2)
300
350
400
Figure 3: Impact of a first-best tax on nitrous oxide and methane emissions, and carbon sequestration 14
3.3.2 Influence on the burden-sharing among EU countries Figure 4 presents the repartition of the reduction in emissions among EU countries. In fact, the relative evolution of each country’s reduction in net emissions reflects the differences in abatement costs and initial emissions. For the net emissions to be reduced, it is necessary that the set of activities in the optimal basis be modified. If not, that means that it is more profitable for a given farmer to pay the tax rather than modifying the optimal basis (in this case, the gross margin decreases linearly with respect to the tax). Thus, for a low level of the tax (from EUR 0 to EUR 15), the incentive to reduce net emissions is too weak to impact significantly total emissions. For the maximum level of the tax examined (EUR 400), EU reduction in net emissions approaches 10 percent of initial emissions from agriculture. 30000
FR
Abatement (,000 tCO2)
IT
25000
GB
GE
PT
GR
IR
SP
BE
ND
DK
LU
20000 15000
10000
5000
10 0 12 0 14 0 16 0 18 0 20 0 22 0 24 0 26 0 28 0 30 0 32 0 34 0 36 0 38 0 40 0
80
60
40
0
20
0
Tax (euro/tCO2)
Figure 4: Burden-sharing among EU Member States
The most striking feature of figure 4 is the large share in the total reduction borne by French and 15
Italian farmers. As far as France is concerned, this result is consistent with its importance in EU agriculture and the low abatement costs found above. The shape of Italian abatement is more interesting, as it reveals a high potential of abatement at a cost ranging from EUR 20 to EUR 35. This abatement is mainly obtained thanks to a reduction in methane emissions through substitutions in animal feeding. Once this abatement is exhausted, Italian emissions remain relatively flat, indicating a strong rigidity in emissions and a nearly-linear decrease of total gross margin. By contrast, emissions from German agriculture appear to be first more expensive to reduce, with a potential of abatement becoming significant for levels of the tax higher than EUR 100.
4
Authorization of afforestation on set-aside land
Accounting for carbon sequestration, which corresponds to the category “Land use, change in land use, and forestry” (LULUCF) in UNFCCC terminology, has been an essential feature of the Kyoto Protocol. This point has also been a major cause of the deadlock of the November 2000 Conference of the Parties in The Hague, as countries did not reach an agreement on a common definition and method of measure of carbon sequestration. It should be noted that carbon sequestration figures reported by each country in IPCC Inventory are not taken into account because of their lack of uniformity and reliability1. Carbon sequestration in trees may also raise issues related to the uncertainty on the future use of wood (Feng et al, 2000). However, if properly accounted for in national inventories, carbon sequestration could significantly reduce the abatement costs associated to a given reduction target, as well improving ambient environmental quality. In this section, we examine the possibility offered to farmers to
1
“Among the problems [regarding LULUCF issues] still prevailing are: (a) lack of uniformity in reporting and varying assumptions among Parties. Some are a logical consequence of different national circumstances and some are a consequence of different methodological approach […] and (b) different definitions of anthropogenic activities, included the differentiation between managed and natural forests” (UNFCCC Secretariat, 1997, page 10)
16
afforest set-aside land as defined in CAP measures. In this case, we assume that subsidy to carbon storage is added to the existing payments associated to actual CAP fixed set-aside programs. As in the former section, we first focus on the dual results to discuss the distribution of abatement costs among farmers and then use the primal approach to derive the implied potentials of reduction. 4.1
Dual approach
The influence of such a measure on the cumulative functions of marginal abatement costs is presented in figure 5. It provides farmers with an additional means to reduce their emissions. As a result, the number of farmers, who are able to achieve a given target of reduction at the same marginal cost, should be higher. Equivalently, achieving a given target should cost less to each farmer. This implies a shift of each cumulative function to the North West. The comparison of figure 5 with figure 2 indicates that this measure would imply an important decrease in the abatement cost for a large number of farmers. For instance, for the highest rate of reduction in emissions (20 percent), the median farmer faces a marginal abatement cost around EUR 175. This represents a decrease of EUR 150 compared to the previous scenario. However, as some farmers do not benefit from this measure because they are not eligible for set-aside payments, such a measure may not be neutral in terms of revenue distribution2. This is particularly true for livestock farmers that have few possibilities to store carbon.
2
For an analysis of the influence of this measure on French agriculture and its potential impacts in terms of revenue distribution, see De Cara and Jayet (2000).
17
1 1% 5% 10% 15% 20%
Percentage in total population
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
100
200 300 400 500 600 700 Marginal abatement cost (euros)
800
900
Figure 5: Cumulative functions of marginal abatement costs when afforestation on set-aside land is allowed.
4.2
Primal approach 4.2.1 Influence on the different sources of GHG emissions
As shown on figure 6, the authorization of afforestation on set-aside land leads to a significant change in the composition of abatement between the different sources. From marginal in the previous scenario, the carbon storage now represents about a third of the total abatement for a value of carbon of EUR 400.
18
Europe
Europe -28
135
StC
N20 CH4
130
-30 -32
Emissions (MtCO2)
Emissions (MtCO2)
125 120 115 110 105
-34 -36 -38 -40 -42
100
-44 -46
95 0
50
100
150 200 250 Taxe (euro/tCO2)
300
350
0
400
50
100
150 200 250 Tax (euros/tCO2)
300
350
400
Figure 6: Impact of a first-best tax on nitrous oxide and methane emissions, and carbon sequestration when afforestation is allowed on set-aside land.
An interesting point also drawn by figure 6 lies in the fact the abatement effort remains oriented towards methane reductions for carbon values lower than EUR 70. Beyond this threshold, carbon sequestration becomes increasingly attractive, whereas the rate of decrease in methane emissions diminishes as low-cost abatement potentials are exhausted. Another noteworthy feature is the interaction between carbon sequestration and nitrous oxide emissions. As area allocated to major crops is lowered, fertilizer use decreases, leading to an additional decrease of 1.2 MtCO2equivalent in N2O emissions (as compared to the previous scenario for a first-best tax value of EUR 400). 4.2.2 Incentives to promote carbon sequestration The impact of the authorization of afforestation on set-aside land differs from one country to another, reflecting the differences in the possibilities of substitution between activities and in the initial endowments in quasi-fixed capital (land and livestock). Except for Belgium, Luxemburg and The Netherlands, this measure modifies significantly the abatement functions and, as a consequence, the potential of reduction that may be obtained through a first-best policy. The abatement differentials between the two scenarios (with and without authorization of 19
afforestation on set-aside land) are shown on figure 7. For carbon values higher than EUR 100, the additional potential of abatement is the most important in France and in Spain. Our results indicate that it is these countries' interests to promote carbon sequestration in international negotiations. In this perspective, they may receive support from Portugal and Denmark, for which the additional abatement is significant comparatively to their initial emissions levels. By contrast, as far as the other major Member States – such as Germany, Great Britain and Italy – are concerned, the additional abatement remains low. 7
1.4 FR SP
5
1
4
0.8
3
0.6
2
0.4
1
0.2
0 0
PT IT GB DK GE
1.2
MtCO2
MtCO2
6
50
100
150 200 250 euro/tC-CO2
300
350
0
400
0
50
100
150 200 250 euro/tC-CO2
300
350
400
Figure 7: Additional abatement allowed by the authorization of afforestation on set aside land.
5
Concluding remarks
Our results show that some potential exists for low-cost abatement in the EU agricultural sector. This is particularly true if strong signals are provided to farmers to increase carbon sequestration, for instance by coupling supply and climate regulation policies. In this case, important reductions can be achieved – most of them occurring in France – at a cost comparable to the abatement costs in other sectors. If not, we show that the potential is limited to about 10 percent of initial emissions even for a high level of the tax of EUR 400 per ton, mainly obtained through methane reductions. For a more reasonable carbon value of EUR 70, we evaluate the European abatement potential to be slightly less than 10 MtCO2, or approximately 4 percent of initial emissions from 20
agriculture. However, the success of a regulation policy aimed at encouraging carbon sequestration requires that an agreement is reached – both among EU countries and in the international negotiations – on the definition of carbon sinks and on a common method of accounting. The last section of this paper highlights the discrepancies that may exist in the EU to this respect. France would benefit the most from measures that favor carbon sequestration. Meanwhile, other major EU countries, such as Germany and Great Britain, would gain little to promote carbon sequestration, while in the meantime having to contribute more to the financing of the CAP. Germany and Great Britain may thus be reluctant to support carbon sequestration in international negotiations.
21
References Babcock, B. A., and Pautsch, G. R. (1999). Relative Efficiency of Sequestering Carbon in Agricultural Soils through Second-Best Market-Based Instruments. Working Paper, FEEM, 90.99, Fundazione Eni Enrico Mattei, Venice, Italy. Balesdent, J. (1995). Stockage de Carbone dans les Sols en Fonction de leur Utilisation. Dossiers de l’Environnement 10 :39-47. Bouwman, A.F. (1989). The Role of Soils and Land Use in the Greenhous Effect, Working Paper, ISRIC, Wageningen Agricultural University, The Netherlands. Burniaux, J.M. (2000). A Multi-Gas Assessment of the Kyoto Protocol. Working Paper, ECO/WP 43, OECD, Paris. De Cara, S and Jayet, P.-A. (2000). Emissions of Greenhouse Gases from Agriculture: The Heterogeneity of Abatement Costs in France. European Review of Agricultural Economics 27(3): 281-303. Feng, H, Zhao, J., and Kling, C. L. (2000). Carbon Sequestration in Agriculture: Value and Implementation. Working Paper, WP-00 256, CARD, Iowa State University. Germon, J.C. and Henault, C. (1995). Processus d’Emissions de Methane et d’Oxydes d’Azote Gazeux par les Sols : Evolution, Quantification, Spatialisation. Dossiers de l’Environnement 10. Hayhoe, K., Jain, A., Pitcher, H., MacCracken, C., Gibbs, M. Wuebbles, D. Harvey, R. and Kruger, D. (1999). Costs of Multigreenhouse Gas Reduction Targets for the USA. Science 286: 905-906. IISD (2000), Summary of the Sixth Conference of the Parties to the UNFCCC. Earth Negotiation Bulletin. http://www.iisd.ca IPCC (1996). IPCC Guidelines for National Greenhous Gas Inventories: Reference Manual. IPCC, Geneva, Switzerland. IPCC (2000). Greenhouse Gas Inventory data. http://www.ghg.unfccc.int Manne , A. S. and Richels, R. G. (2000). A Multi-Gas Approach to Climate Policy – with and without GWP. Working Paper, EMF-19, Washington, DC. Newell, R. and Stavins, R. (2000). Climate Change and Forest Sinks: Factors Affecting the Cost of Carbon Sequestration. Journal of Environmental Economics and Management 40(3):211 Plantinga, A., Mauldin, T. and Miller, D.J. (1999). An Econometric Analysis of the Costs of Sequestering Carbon in Forests. American Journal of Agricultural Economics 81(4):812-824. Reilly, J., Prinn, J. R., Harnisch, J., Fitzmaurice, H., Jacoby, D., Kicklighter, D, Meillo, J., Stone, P., Sokolov, A., and Wang, C. (1999). Multi-Gas Assessment of the Kyoto Protocol. Nature 401: 905-906. Sauvant, D., Giger-Reverdin, S., Tregaro, Y (1996). Qunatification de la Production de Methane 22
par les Ruminats. Working Paper, Laboratoire de Nutrition et d’Alimentation INRA de l’INA PG, INA PG, Paris, France. Schneider, U. A. (2000). Agricultural Sector Analysis on Greenhouse Gas Emission Mitigation in the United States. PhD thesis, Texas, A7M University. Simons, B., Burrell, A., Oskam, A., Peerlings, J., Slangen, L. (1994). CO2 Emissions from Dutch Agriculture and Agribusiness: Method and Analysis. Working Paper, Wageningen Economic Papers, Wageningen Agricultural University, The Netherlands. UNFCCC Secretariat (1997). Methodological Issues: Synthesis of Information from National Communications of Annex I Parties on Sources and Sinks in the Land-Use and Forestry Sector. Technical Paper FCCC/TP/1997/15, UNFCCC.
23
