University of Groningen Faculty of Economics and Business MSc International Economics and Business
Master thesis
Explanatory economic assessment of a potential emission
trading system for EU member states in agriculture
Name Student: Iris van de Vegte Student ID number: S2783606
Student email: i.van.de.vegte@student.rug.nl Date: 10-02-2017
2 Abstract
This paper examines the impact of policy decisions regarding the agricultural sector as part of the burden sharing decision in the EU. Using a panel data approach, the emission reduction potential for old and new member states is analyzed. For the agricultural sector the
conclusions are from an economic perspective in favor of the tradable emission allowance system in order to reduce emissions in a cost-effective way. The added value of this thesis is to investigate the difference in capital intensity for old and new member states as an addition to the literature on agricultural emission reduction costs in the EU.
3 List of abbreviations
EU European Union
GDP Gross Domestic Product
EEA European Environment Agency ETS Emission Trading System ESD Effort Sharing Decision BSA Burden Sharing Agreement DOP Domestic Offset Projects GHG Greenhouse Gases
MAC Marginal Abatement Cost CAP Common Agricultural Policy CO2 Carbon dioxide
CH4 Methane
N2O Nitrous oxide
MtCO2eq Metric tons of carbon dioxide equivalent in millions FADN Farm Accountancy Data Network
4 List of tables and figures
Figures
Figure 1: MAC curves and emission trading………..18
Tables Table 1: (Agricultural) emissions and targets as part of the non-ETS for old and new member states..…...………...…...………...…...………....…...…………...15
Table 2: List of the variables………..…...………...28
Table 3 Descriptive statistics ……….………....…...………...…...…………..31
Table 4: Estimated regression results with random effects …………..……….………..32
5 Table of Contents
Abstract ... 2
List of abbreviations ... 3
List of tables and figures ... 4
I. Introduction ... 6
II. Problem description and literature review ... 9
Policy background ... 10
Current ETS allowance price ... 11
Agricultural focus ... 13
Emissions from agriculture ... 14
Achieving cost-effectiveness ... 17
Potential welfare gains and economic implications ... 19
The research question ... 21
Hypotheses ... 22
III. Data and method ... 26
Data specification ... 26
Dependent variable ... 27
Independent variables ... 27
Methodology ... 29
Quality of the data ... 29
Model description ... 30
Descriptive statistics ... 31
IV. Empirical results ... 32
Results ... 32
Robustness check ... 33
V. Conclusion and discussion ... 36
6 I. Introduction
Climate change is a hot topic these days. For policymakers, it is a challenge to deal with environmental issues. More policy incentives are needed to cope with the climate problems of the current era and encourage pro-environmental behavior. This thesis focuses on the cost-effectiveness potential of reducing emissions in Europe. There is a distinction made between the original fifteen EU member states (called the ‘old member states’) and the 2004 onward expansion, described as ‘new member states’. This distinction has been done in order to indicate the difference in farm capital intensity of old and new member states which might give an indication of energy intensity in favor of an emission trading system.
The evaluation follows the agricultural sector as part of the non- emission trading system (non-ETS). The emissions in agriculture have negative externalities on the market, meaning the external costs of agricultural production on society such as the carbon pollution. Government interventions are needed to reduce these negative externalities and at the same time, firms need incentives to reduce their emissions in a cost-effective way. For sectors who are part of the European Emission Trading System (ETS) these incentives are available in the form of credits. Credits are tradable across Europe at a certain price. The European ETS works on the so called 'cap and trade' principle (European Commission, 2016c) also known as permit or allowances trading. This ‘cap’ is set on the total amount of greenhouse gases (GHG) that can be emitted by energy production plants who are part of the ETS. The cap is reduced over time, so total emissions have to fall. These power plants have the option to buy credits and trade with other power plants. In the end, these energy production plants must have collected the right number of credits to cover their emissions. If a company reduces the emissions, it can sell credits to other installations who are short on these credits to cover their emissions. This is how supply and demand is established and the market of emission permits arises under the ETS (European Commission, 2016b).
However, there is also an emission reduction potential in other (non-industrial) sectors but which are not covered by an emission trading system within the EU. These sectors do not have the same incentives to reduce emissions as the sectors belonging to the ETS do.
7 establishment of the member state targets by the European Commission. These specific
targets are set for the non-ETS sectors to contribute to the overall emission reduction targets of member states for 2020 and 2030. But the implementation to reduce emissions in these sectors is entirely up to the member states themselves. This makes it a challenge for member states to lift off emission reduction projects per sector. To support GHG abatement measures (also known as reduction measures), the so-called Domestic Offset (DO) project scheme has been introduced by EU policymakers as an attempt to stimulate emission reductions in the sectors not covered by the ETS. However, these projects are not yet implemented, and scientists argue about the contribution of such an extra policy option in comparison with the already existing ETS.
In this thesis, research is done to explain the potentials for an emission trade system in the agricultural sector which will be determined by the differences in emissions reduction costs across member states. In this study, the distinction is made between new (high income) and old (low income) member states to map the differences in emission reduction costs by means of the capital intensity differences between these two groups. A panel dataset (with macro data) is used for the capital intensity variable as main determinant on livestock (which is the most capital intensive form of farming) farm output of 27 EU member states between 2007 and 2013. The added value of this thesis is to investigate the difference in capital
intensity proxy for old and new member states, as an addition to the literature on the variation in agricultural emission reduction costs across EU member states.
A theoretical background based on literature from De Cara and Jayet (2011) on this matter is presented to support a potential cap and trade system for agriculture. On basis of the theoretical background and more additional literature, this study is in favor of an emission trade system for the agricultural sector from a cost-effective perspective. The research question that will be answered is: ‘To what extent can there be gains from an emission trade
system for new and old member states if there are capital intensity differences in agriculture?’ Essentially, this study is twofold. First, the study theoretically explains the potential
gains of an emission trade system in agriculture from a cost-effective perspective. Secondly, in the empirical analysis, the capital intensity differences in agriculture for new and old member states are investigated to indicated the energy intensity differences, in order to ascertain if an emission trade system between member states would be cost-effective.
8 agriculture in order to reduce emissions at an European level. The potential (loss of) cost-effectiveness of implementing the policies across Europe, by means of a theoretical model called the marginal abatement cost curve (MAC curve), is presented in this section as well. In the third chapter, the data collection process and methodology is described. Then, an
9 II. Problem description and literature review
In the fall of 2016 the European Commission decided to set stricter GHG reduction targets. All EU member states agreed on setting a binding domestic emission reduction target of at least 40% by 2030, compared to 1990. All member states across all sectors in the EU should contribute to meet these emission reduction targets. To do so in a cost-effective manner, the industrial and power sectors covered by the EU ETS should reduce emissions by 43% in 2030 compared to 2005. The non-ETS sectors should reduce their emissions by 30% in 2030 compared to 2005 (European Commission, 2017b). According to the latest GHG inventories by the EEA, agricultural emissions represent about 10% of the total EU emissions (as part of the non-ETS) (European Environment Agency, 2010). The recent European decisions set various objectives for GHG emission reductions (European Commission, 2008b). To meet these objectives, the European Commission has defined a strategy. On the one hand, GHG emissions from large-scale emitters, mostly in the industry and the energy sectors, are currently covered by a cap and trade system known as the ETS (European Commission, 2003). On the other hand, emissions from for instance, the transport, construction, and agricultural sectors, which are much less concentrated, are not subject to emission trading. The decision to not cover the former sectors with an emission trade system was however accompanied with a burden sharing agreement (BSA), which sets member state specific targets for non-ETS emissions (European Commission, 2009a). The current BSA, also known as the European Effort Sharing Decision (ESD) (406/2009/EC), sets targets per individual member state and an overall 10% GHG reduction target for 2020 (compared to 2005 levels), for European GHG that are not covered by the ETS. The European effort to meet this target is shared among member states in such a way that GDP effects are distributed in a fair and equitable manner according to the European Commission (European Commission, 2009a). This is considered to be the case under the following effort sharing principles:
- The member state that had the lowest level of GDP per capita in 2005 is allowed to increase its emissions by 20% in 2020 (compared to its 2005 level);
- The reduction targets for the member states that had the highest levels of GDP per capita in 2005 are set at 20% below those 2005 levels;
- The reduction targets for member states that, had a GDP level per capita equal to the EU average in 2005, are set at the average EU reduction target;
10 the minimum and average targets. For those with above average level of GDP per capita, the function is based on the maximum and average targets (Verdonk and Hof, 2013). However, higher targets for high income member states compared to lower targets for low income member states also mean higher costs in achieving these targets. To address this difference between high and low income member states, this demands a flexible system in which member states can reduce emissions as a collective effort. The high income countries are recognized as old member states (the original EU15) while new member states (from the 2004 onward EU expansion) are recognized as low income countries (European Commission, 2016d).
In contrast to sectors in the EU ETS, which are regulated at EU level, it is the responsibility of member states to define and implement national policies and measures to limit emissions from the sectors covered by the ESD (European commission, 2016d). Member states have significant differences in economic and investment capacity as well as in reduction potential. The application of cost-effectiveness as a criterion for the distribution of efforts would lead to considerable variations in the necessary national economic efforts. However, setting targets that take fairness and solidarity into account may result in large differences in the costs of reducing emissions between member states. These differences are due to a wide variety in the economic capacity of EU member states in terms of emission reductions costs. The existing flexibility within the BSA needs to be increased to allow the achievement of cost-efficient emission reductions across the EU (European Commission, 2009a). This calls for a possible option of tradable emission allocations in the non-ETS sector and collective efforts of the EU member states to meet the targets in the same way this is done in the current ETS sectors.
Policy background
11 with a solid plan for emission reduction in a non-ETS sector (e.g. agriculture). The developers of such projects (e.g. farm owners) can receive emission credits for the potential emission reduction. These rights have a certain value and can be sold to domestic entities who want to compensate their emissions. However, such projects are still in their infancy. According to Oikonomou et al. (2012) DOP’s could provide an extra stimulation for farm owners to promote energy efficiency investments (Oikonomou et al., 2012). Van der Gaast et al. (2013) described in their paper to link the DOP’s with the ETS market. According to Van der Gaast et al. (2013) this option has some advantages. The authors think that the DOP mechanism stimulates development of low-carbon energy technologies within the EU and could unlock enormous amounts of CO2 savings in non-ETS sectors. DO projects could broaden the scope
for ETS installations to comply with their annual targets and the economic value of GHG credits reduces the need for government grants outside the ETS (Van der Gaast et al., 2013). Current ETS allowance price
In this study, some attention is paid to the current allowance price of the ETS to consider the possibility of a non-ETS sector joining this system. The EU ETS allowance price dropped from approximately 22€/tCO2 in 2008to approximately 5€/tCO2 in 2016. The cause of the
sharp price drop can be explained by the economic crisis in 2008 according to Aldy and Stavins (2012). There was a lower demand for allowances due to reduced output in the energy-intensive sectors (Aldy and Stavins, 2012). However, the economic recession might not be the only explanation for low allowances prices. According to Frankhauser et al. (2010) there are also overlapping climate policies which might not always have the desired effect. In their paper, the authors review the implications for the carbon price of combining cap and trade with other policy instruments. Frankhauser et al. (2010) argue that policymakers attempt to control carbon prices (price of those who emits the CO2) by multiple policies is often
inefficient and ineffective. It can have adverse consequences, such as undermining the carbon price and increase mitigation costs. A stable carbon price remains a desirable policy objective according to Frankhauser et al. (2010).
12 The market stability reserve is introduced by the European Commission to address the surplus of allowances in order to push up the price of emission rights (European Commission, 2017a). This system takes a portion of the allowances of the market and put it in the reserves. The European commission (2017a) also state that efforts to address the market imbalance would help to faster reduce the assigned annual cap of member states. This mechanism will be introduced in 2018 and the reserve will start operating in 2019 (European commission, 2017a).
Knopf et al. (2014) state that no significant increase in the allowance price is expected before 2020. Knopf et al. (2014) discuss that reforms are needed and policymakers need to come up with mechanisms that reduces uncertainty and stabilize expectations of participants in the EU ETS. The authors argue that the market stability reserve will not address the problem of dynamic efficiency nor the problem of overlapping policies that undermine the performance of the EU ETS (Knopf et al., 2014). The analysis of Knopf et al. (2014) show that a comprehensive reform addressing several aspects of carbon pricing is needed. One of the reforms, which is especially interesting for this thesis is that Knopf et al. (2014) are in favor of expanding the EU ETS to other sectors, such as agriculture, transport or construction sector. This flexibility is needed to generate the best results in terms of incentives investments in low carbon-technology in the most cost-effective way.
13 Agricultural focus
The 2008 EU ETS directive (based on the 2003 directive) signals that an extension to the agricultural sector should not be considered due to concerns regarding the accuracy of
monitoring, reporting and verification (European Commission, 2003; European Commission, 2008a). However, agriculture is a significant contributor of GHG emissions. In 2004, methane (CH4) and nitrous oxide (N2O) emissions from agriculture account for approximately 13.5%
of global GHG emissions (IPCC, 2007). Agricultural emissions result from the activities of a large number of small and medium scale emitters across Europe which lead to a range of differences in abatement costs due to the output differences within and between new and old member states. Such heterogeneities have important consequences on the design of cost-effective mitigation policies (De Cara et al., 2005). Studies have indicated the emission mitigation potential in agriculture, such as the study done by Höglund-Isaksson et al. (2012) where the authors describe the development of emission scenarios of non-CO2 emissions in
the European union (Höglund-Isaksson et al., 2012). They show that it is possible to cut greenhouse gases to half of the size by 2020 compared to 2005. A study done by Perez Domiguez et al., (2009), where the authors focus on the application of a permit trade system for emission reduction in agriculture, sheds light on the burden sharing agreement boundaries for the agricultural sector based on economic efficiency (Perez Domiguez et al., 2009). These studies are all in favor of a market based system like the ETS to be applicable for the non-ETS as well.
However, there is a gap in the literature concerning the latest policy development, such as the DOP’s in agriculture and linking it with the cost-effectiveness potential under a cap and trade system for the non-ETS sector, in particular agriculture. Additional literature is needed to underline the potentiality of an emission trading system for agriculture from a
cost-effective perspective to support effort sharing between new and old member states. Trading credits or also known as permits (one emission credit/permit is considered equivalent to one metric ton of carbon dioxide emissions) across member states borders might lead to the most efficient target achievement in every member state (if the agricultural sector will be included). This is because if domestic projects in agriculture decrease overall emissions, a certain
14 emission reduction take place where costs are lowest- could be most cost-efficient. Because when there exists a market of credits which is made attractive, acceptable and has a certain value among EU member states, in theory these countries could maximize the emission reduction at lowest costs.
Emissions from agriculture
In this section the distribution of the non- ETS emissions and agricultural emissions will be outlined. According to De Cara and Jayet (2011) four countries (France, Germany, United Kingdom, and Spain) account for more than half of the total European agricultural emissions (De Cara and Jayet, 2011). These countries are all indicated as old member states, so the total emissions of the EU15 are much higher than the total emissions of the new member states. This is shown in column 1 of table 1, where old member states have a total of non-ETS emissions in 2005 of 2486 MtCO2eq. MtCO2eq are the metric tons of carbon dioxide equivalent in millions. The total non-ETS emissions in the new member states is much smaller, namely 506 MtCO2eq, as presented in row 2 of the first column. This variety in emissions has to do with the relative size and importance of agriculture in the different countries. The wide diversity across member states in terms of GDP per capita and expected growth involves large differences among member states in terms of (total and per capita) GHG emissions and the corresponding reduction targets (EEA, 2010).
In the second column, it is shown that agricultural emissions are much higher in old member states compared to new member states. Old member states have approximately 407 MtCO2eq while new member states only have 96 MtCO2eq agricultural emissions as part of the non-ETS. This is shown in respectively the first and second row of the second column of table 1.
15 In the last column of table 1, the abatement target is indicated for the old and new member states. The range from targets is between 20% and -20% set by the European Commission (2009b). A negative target means that the member states may increase its non-ETS emissions until 2020. This counts for almost all new member states except Cyprus. There are however large differences in the targets between new member states, Slovenia and Malta for example can only increase emissions respectively 4% and 5% while Romania and Bulgaria can increase emissions respectively 19% and 20% (European commission, 2009b). The total is shown in column 4 of table 1, where on average old member states have a reduction target of 14%, while new member states have on average an increase of emissions of 14%.
Member states Non – ETS emissions (total MtCO2eq) (2005 emissions) Agricultural emissions (total MtCO2eq) (2005 emissions) Share of agricultural emissions in non-ETS (% average of the member states) Non – ETS abatement target (% average of the member states) (2020/2005) Old member states EU15 2486 407 16 14 New member states 506 96 19 -14 Total EU27 2992 503 17 9
Table 1 (Agricultural) emissions and targets as part of the non-ETS for old and new member states
Notes table: In the second column, the total emissions are the methane (CH4) and nitrous
oxide (N2O) emissions from agriculture. In the fourth column, a negative target means you
can increase the emissions in the non-ETS sectors. Data for the table is used from De Cara and Jayet (2011) and European Commission (2009b).
The EU Common Agricultural Policy (CAP) tries to reform agricultural production in the current decade towards 2020 by decreasing the production capacity of farms across EU member states. The EU CAP targets the environmental impacts of agricultural production via policy reforms and stimulate more environmental friendly farming (European commission, 2016e). Intensive agricultural activities in the EU contribute to emissions of greenhouse gases (CO2) but also methane (CH4) and nitrous oxide (N2O) (IPCC, 2007). Agricultural activities
are among the major contributors of the total EU GHG (European Environment Agency, 2002).
16 According to Olesen (2006) mitigation strategies will be required within the agricultural sector to comply with the reduction targets of the European commission (Olesen, 2006). Climate change affects agriculture differently across countries. Across the EU member states there will be large differences in adapting to the proposed policies from the European
Commission (Parry et al., 2004). However, this is not only dependent on the political system of the particular country, but also on the climate conditions which effect soil and land use (Bouma et al., 1998). According to Olesen and Bindi (2002) these differences are expected to greatly influence the responsiveness to climate change. Furthermore, emissions occur
differently across farms, making it hard to evaluate cost-effectiveness across member states due to different techniques available and the corresponding cost of implementing these techniques.
In recent years, greenhouse gas emissions from agriculture have been influenced by a number of factors, such as economic trends and farm management practices, but also the reforms in the CAP (Eurostat, 2015). These previous mentioned influences had their impact on the agricultural practices (and most importantly the corresponding emission limits) across the EU. The reduction in agricultural emissions of greenhouse gases can be achieved via more efficient and climate friendly farming practices (Eurostat, 2015). The Eurostat database (2015) is conducted in this paper as well to indicate the greenhouse gas potential across member states. Research done by Eurostat (2015) states that the EU member states with the largest agricultural sector tend to account for the highest greenhouse gas emissions from agriculture production (Eurostat, 2015). France and Germany together contributed just over one third of the EU28 greenhouse gas emissions from agriculture in 2012. It can be concluded that most countries in the old member states (original EU15) accounts for the highest
emission share in agriculture (Eurostat, 2015). This conclusion is in line with the database from de Cara and Jayet (2011).
According to Van Doorslaer et al. (2015) the emission decline has slowed and
European emission reductions are expected to stagnate in the (near) future (Van Doorslaer et al., 2015). The authors state that the member states emission reductions slowed between 2001 and 2011 compared to earlier decreases in the 90’s. The productivity increase is one of the main factors attributed to stagnation next to the development of agricultural and
environmental policies (Van Doorslaer et al., 2015).
17 have a certain impact on both old and new member states and influence the emission
reduction potential for the EU member states. The comparison between economic and environmental factors on farm performance is investigated and linked to the European
policies. By making this distinction between old and new member states the adaptive capacity of new and old member states on policy reforms will be indicated.
The wide diversity across member states in terms of GDP and expected (agricultural) growth involves large differences among member states in terms of the emission reduction potential in agriculture (European Environment Agency, 2010). The share of agriculture in non-ETS emissions varies greatly and according to De Cara and Jayet (2011) it is higher on average in the new member states than in the rest of the EU. De Cara and Jayet (2011) argue that if there was a trade system based on emission allowances, new member states will mostly sell permits while old member states buy these permits. This is because old member states have larger agricultural emission reduction targets and also higher abatement costs (De Cara and Jayet, 2011). This means lower abatement costs are expected according to De Cara and Jayet (2011) in new member states, because of the lower targets set under the BSA. This is in contrast with the old member states and higher cost of emission reduction is expected which, in the end, is less cost-effective.
Achieving cost-effectiveness
In this section the economic and environmental effectiveness of implementing an emission trade system for the agricultural sector as part of the non-ETS will be outlined by means of an illustration.
A marginal abatement cost (MAC) curve is defined as “...a graph that indicates the cost in a currency per ton of CO2, associated with the last unit (the marginal cost) of
emissions abatement for varying amounts of emission reduction” (Kesicki and Strachan, 2011). The advantage of the MAC curve is that it shows the emissions marginal abatement cost for every extra unit of emission reduction. This curve can be used by policymakers when implementing cost-effective market based policies for emission reduction. The total cost of achieving the emissions reductions needed are minimized, such that the abatement costs are equalized at the margin between countries (Kesicki and Strachan, 2011). The graphical illustration of the MAC curve with a cap and trade system is based on literature from Tietenberg and Lewis (2012).
18
Figure 1 MAC curves and emission trading based on literature from Tietenberg and Lewis (2012)
The total abatement requirement in MtCO2eq for the two countries is the width of the
19 as explained earlier. Trading reduces overall abatement costs by area D + E. This area can be interpreted as the measure of the cost saving by using economic instruments such as a cap and trade system (Tietenberg and Lewis, 2012).
However, some assumptions and limitations of the MAC curve and emission trading need to be taken into account. First of all, the allocation of allowances is determined by the European Commission. This is done via auctioning the allowances (creating government revenue) or by giving allowances away for free (also known as ‘grandfathering’) to member states. According to the European Commission, a surplus of emissions has been built up since 2009 (European Commission, 2017a). This surplus undermines the cost-effectiveness. When prices are too low due to oversupply this reduces incentives to invest in more climate friendly technologies as stated earlier by Nordhaus (2011). As a short-term solution, the European Commission postponed the auctioning of emissions until 2019 – 2020. As mentioned before, the market stability reserve is introduced as a long-term solution were unallocated allowances will also be transferred to the reserve. The exact abatement costs in agriculture for old and new member states is unknown. The MAC curves are not equalized across countries and thus the exact costs effectiveness is hard to measure.
As the information for the exact CO2 marginal abatement cost per hectare in
agriculture is currently not available, a different approach is used to indicate the cost potential for old and new member states. There will be looked at the capital intensity of farms to give an indication of energy intensity in agriculture for old and new member states. If there is indeed a difference as the literature from De Cara and Jayet (2011) suggests, this gives an indication that the MAC curve differs between old and new member states, and potential gains from trade can be obtained. If new member states are allowed to increase emissions, this speaks even more in favor of a trade system. New member states will mostly be selling
permits, while old member states will be buying these permits. This statement is also in line with literature from Perez-Dominguez et al. (2009) were the authors state that western
European countries are mostly permit buyers and eastern European countries permit providers (due to the assumed lower expected abatement costs).
Potential welfare gains and economic implications
20 welfare gains can be achieved if more flexibility under the BSA was given to member states. Access to a market based system for the member states has a positive impact on welfare as such a system might limit the emission reduction cost. The cost reductions based on the suggested policy change towards a market based system in this section improve efficiency across member states. This way, the targets can be met at much lower cost. This is
investigated by research from Ranson and Stavins (2015). The authors state that each country which accepts allowances and participates in trade can be advantageous because of the
welfare gains from this exchange between countries with different comparative advantages. If policy makers have the goal to maximize welfare, then these benefits would affect policy decisions in favor of linkages between cap and trade systems. Furthermore, the authors argue that the most important reason for linking systems is the increase in cost-effectiveness that results from the allocation of abatement costs between systems with different MAC curves (Ranson and Stavins, 2015).
The economic implications of a trading situation will be discussed below. A rough estimation based on literature from Perez-Dominguez et al. (2009) is given with regard to the situation when a trading scheme for agriculture is introduced for EU member states.
According to Perez-Dominguez et al. (2009) when emission trade is introduced, the sales of permits go up to 8.2 billion MtCO2eq (1 permit is 1 MtCO2eq) within 5 years. The authors also calculated the potential permit sales when permit trading is restricted and only
agricultural producers within a member state are allowed to trade with each other. In the last case, only 3.2 billion MtCO2eq is traded, but the authors state that the permit price
convergence between regions within a member state will be less smooth and abatement costs increase. Unrestricted permit trade across member states is thus more cost-effective according to Perez-Dominguez et al. (2009). The authors calculated the range of the permit prices when there are permit trade possibilities in agriculture across member states. Within the EU27 the permit price would range between 68 euro - 82 euro (in a hypothetical world without
implementation costs). So, the rough estimation of the economic profits when there is a trade situation of permits across EU member states range from 558 billion euro up to 672 billion euro (68 euro * 8.2 billion MtCO2eq and 82 euro * 8.2 billion MtCO2eq). This estimation should be interpreted with some assumptions. The rough estimation is derived from the
21 economic profits. The authors assumed that potential abatement costs are higher than the permit price. Exact abatement costs for old and new member states are currently not (yet) available. As explained earlier the current permit price is much lower than the price Perez-Dominguez et al. (2009) suggests. Therefore, more research is needed for the potentials and costs of a trade system when permit prices are lower, but this is not discussed in this paper. Lastly, the authors conclude that there are potential economic welfare gains of a cap and trade system compared to other less flexible emission abatement possibilities. According to Perez-Dominguez et al. (2009) consumers will benefit because food prices will be kept low. This is due to the cost-effectiveness potential of an international trade system ‘at work’. Otherwise, if countries have to reduce their emissions domestically without the flexibility of an (across border) trading option, higher abatement costs are the result of this approach, which is reflected in higher food prices for consumers, inducing a welfare loss. Moreover, the authors state that additional revenues are expected for agricultural producers through the trading of emission permits.
The research question
The impact of the different policy guidelines on agricultural GHG emissions over time is the combination of measures influencing farm input and outputs. If member states can buy and sell reduction allowances in agriculture with other member states, cost-effectiveness can be achieved. In this case, emission reduction can be achieved at the location where it is the cheapest to reduce emissions. The extra costs member states make when there is not a trading system do not arise anymore. In the absence of information on the exact emission reduction costs of farms included in the FADN database from the European Commission (European Commission, 2016a), the capital intensity in agriculture is investigated instead. The capital intensity variable is used to (indirectly) test if there is enough evidence to assume that there are differences in energy intensity in agriculture. This provides (indirect) evidence for
marginal cost differences for old and new member states in favor of an emission trade system. The main research question of the paper is: To what extent are there gains from an emission trade system for new and old member states if there are capital intensity differences in agriculture?
22 Hypotheses
In this section, the hypotheses are formulated based on and the objective of the paper. Via an extensive literature review regarding the (expected) differences in the European emission reduction costs in agriculture, the hypotheses can be formulated.
Vanschoenwinkel et al. (2016) applied a case study in eastern European countries (the new member states) to better understand the differences between eastern and western
European countries on the impact and associated costs of climate change. They argue that new member states suffer from agricultural climate change policies due to their low adaptive capacity (both economically and politically). The authors are a big supporter of policy
changes to give new member states better options in adapting to agricultural climate response (Van schoenwinkel et al., 2016). This would mean policy makers need to rethink policy strategies regarding incentives for member states to reduce emissions in a cost-effective manner.
According to Grosejan et al. 2016 there is large variety in the emission reduction potential and costs across emission sources due to dissimilar biological processes. Differences in farms size can also lead to varying mitigation potentials and costs. For instance, some emission reduction options require larger investments and these options might be more accessible to large farms because of economies of scale (Grosejan, et al., 2016). However, farm size is not considered in this thesis to be part of the empirical analysis, but instead the focus will be on the capital determinants in agriculture to explain the difference in livestock farm outputs for new and old member states.
In addition to the heterogeneity of mitigation potential across emission sources, literature from De Cara and Jayet (2011) and Perez-Dominguez and Fellmann (2015) stated that there are different MAC curves across member states, with the lowest abatement costs generally located in new member states (De Cara and Jayet, 2011; Perez-Dominguez and Fellmann, 2015). The studies also show that it does not cost the same for all producers within and across member states to reduce emissions. MAC differs a lot between agricultural
activities and regional farming systems (De Cara and Jayet, 2011; Perez-Dominguez and Fellmann, 2015). However, the exact results of the studies and therefore the exact abatement costs of farms are not accessible. Nevertheless, both studies are clearly in favor of redesigning the current climate policies to deal with these cost differences and reduce GHG emission in the agricultural sector without unnecessarily high costs.
23 differences in capital intensity of farms for old and new member states will be investigated. If capital intensities are relatively high, it means that emission reduction costs are also relatively high. Literature on this matter comes from Cole and Elliott (2005) who conclude that the more intensively a particular sector make use of capital, typically means the more pollution intensive this sector is. The authors state that capital abundant countries are typically those with some of the strictest environmental regulations and are mainly located in the more prosperous member states, indicated as the original EU15 (Cole and Elliott, 2005). The study of Cole and Elliott (2005) also found that in the capital abundant country, the capital/labour ratio will be higher, indicating a more capital intensive production process instead of labour intensive. Capital-intensive production processes appear to generate more pollution per unit of output than labor intensive processes. Overall, a capital-intensive production process is an indication for an energy intensive production processes, according to Mani and Wheeler (1998) and Gerlagh and Mathys (2011). Both studies find clear evidence of complementarities between capital and energy in the production function.
As stated by the European Commission (2014) based on FADN data, the old member states (e.g. Denmark and the Netherlands) are characterized by higher average total assets values of farms (as indicator of capital intensity), while new member states (e.g. Romania and Bulgaria) have much lower average total asset values. The total output in the EU15 is also much higher than total output in the countries not part of the EU15. According to the European Commission (2014) this can be explained by the higher returns on assets in old member states than in new member states. The higher return on assets in the old member states gives an indication on the higher capital intensity in these countries. Thus, old member states are more oriented towards high capital intensive farming while new member states are oriented towards less capital intensive farming (European Commission, 2014). The type of farming is relevant to indicate the capital intensity according to the European Commission (2014). Typically, diary (cow’s milk) and granivore (pigs and poultry), both livestock farming, are characterized by high averages total assets compared to field crops in the EU. Thus, farms focusing on livestock are more capital oriented while farms focusing on crops are less capital oriented. Therefore, the type of farming can be an important determinant to give the best indication of whether there are emission reduction cost differences between member states. The agricultural variable will be specified by the proxy of the output of livestock farming in the empirical analysis.
24 new member states under the BSA. To give an indication on the energy intensity differences of European farms in this study, the capital intensity of livestock farming (which is the more capital intensive form of farming according to the European Commission (2012)) will be investigated. If there is enough evidence to accept that there are significant differences in capital intensity on livestock farms output, this will give an indication of abatement cost differences for old and new member states. Therefore, the hypotheses of the thesis are as follows:
Hypothesis 1. There can be gains from an emission trade system for old and new member states because there are capital intensity differences in agriculture
Hypothesis 2. The capital intensity of livestock farms is higher in old member states than in new member states
Literature on using a capital intensity variable on basis of the FADN dataset (European Commission, 2016a) comes from Curtiss (2000), Davidova et al. (2003), Lissitsa and Odenlin (2005), Iraizoz et al. (2007) and Dos Santos (2013). The capital intensity measure included in this study is according to previous mentioned literature. Capital intensity was approximated by the quantity of depreciation per annual work unit (Depreciation/AWU), with higher values indicating that there is more capital available per employed worker. According to above literature, if there are greater capital intensities per worker, this will indicate higher total output but also a more energy intensive production process. If there is indeed evidence that countries in the EU15 have a more capital intensive production process, this indicates higher emission reduction costs for these countries. With regard to the use of the depreciation variable, literature from Petrick and Kloss (2012) state that fixed capital inputs are
approximated by depreciation of capital assets estimated at replacement value in euros and therefore this variable is best to indicate fixed capital over the years (Petrick and Kloss, 2012). The depreciation variable includes for example, farm buildings and fixed equipment, land improvements and machinery & equipment (European Commission, 2016a). This variable, together with the AWU variable are derived from the FADN dataset, however only the calculated capital intensity variable will be included in the final regression.
25 such as machines and buildings but also the circulating capital (current assets)). These
26 III. Data and method
In this chapter the specification of the final sample is outlined. The dependent and
independent variables are specified. The methodology of the study is described in this chapter as well.
Data specification
The specification of the data sources, data collection process and final sample is specified in this paragraph. A regression model is used for the econometric part of the thesis. A linear regression model is performed to test the hypotheses. The information necessary to perform the regressions was taken from the European Commission database specialized in agricultural and rural development. This database is called the Farm Accounting Data Network (FADN), covering all EU member states from period 2004 until 2013 (European Commission, 2016a). FADN is an unbalanced panel because not each year the same farms are represented in the data. An advantage of using the FADN is that it includes information on all kinds of farming activities that cannot be found in other EU wide datasets. However, some disadvantages of the FADN dataset need to be outlined as well. Due to privacy concerns, only total country
averages of farm input and output measures were available in the FADN dataset, limiting the number of observations. With regard to the environmental assessments, no variables were included in the dataset about energy intensity or CO2 emissions. Furthermore, the variables are expressed in economic terms only, e.g. in euro instead of kilograms, which might limit the interpretation of the results. Regarding the direct abatement of farms, this could not directly be derived from the FADN dataset, as the dataset does not contain any data on this matter as explained before. The intensity of capital is therefore used to indirectly indicate the amount of capital per unit of labour differences across member states. This gives an indication (no more) on environmental performance of farms.
27 Netherlands, Portugal, Spain, Sweden, United Kingdom (Eurostat, 2014). Together with the new EU member states who joined in 2004 making it the EU15 plus Cyprus, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia, which is together the EU25. Finally, the two member states who joined in 2007 are also added to the final sample, namely Bulgaria and Romania, and are also indicated as new member states. The final sample covers 27 member states from the period of 2007 till 2013.
The FADN dataset provides only information of the economic situation on average per member state rather than per company or farm in this case. Therefore, averages on the
different economic indicators per member state are used in the sample to analyze and measure the impact for policy implementation in agriculture. In this way, the cost-effectiveness per member state can be indicated by looking at the average farm output at European level. Limitations on this approach is presented in the last chapter of the paper. The FADN data is harmonized across all EU member states and can be used for the evaluation of, for example, agricultural policy as well as in monitoring the farm income levels across member states. To avoid biases caused by fluctuations, e.g. in production (due to bad weather) or in input or output prices, the FADN data is corrected for price fluctuations (inflation) and measured in real values to make it comparable across member states over the years. Aggregated yearly data is therefore accessible in the public database from FADN (European Commission, 2016a).
Dependent variable
The total livestock revenue is an output measure for agricultural production measured in euros. This variable is the dependent variable to test the hypotheses and specified in the regression as Livestock Output. The variable specification is based on the FADN dataset, which is the sales in euros from livestock production (quines, cattle, sheep, goats, pigs, poultry and other animals), change in livestock value and sales from animal production (milk, milk products, ewes, wool, eggs). The specification of a certain type of farming is needed to appropriately compare farms across member states. This way, a more homogenous set of products will be compared.
Independent variables
28 The first variable is the capital intensity variable derived from the variables depreciation divided by the annual work unit (Depreciation/AWU) in Stata. AWU corresponds to work done by one employee working on a full time yearly basis. The Capital intensity variable therefore corresponds to the quantity of capital per AWU. The (final) Capital Intensity variable will be included in the regression. This variable is expressed in euros. The second variable is the dummy variable EU15 indicating if the country is part of the EU15 or not. The variable takes a value of 1 if the country is part of the EU15 and 0 if the country is not part of the EU15. Furthermore, an interaction term is included to test the hypothesis if old member states have higher capital intensity in livestock farms compared to new member states. The variable is indicated as EU15_Capital Intensity and derived from the first two independent variables (the continuous variable; capital intensity and the dummy variable; EU15). An overview of all the variables is presented in table 2. In this table, all variables are listed together with the corresponding measurement used in Stata plus a brief explanation. The descriptive statistics with the mean and standard deviations are presented in table 3 in the at the end of this section.
Table 2 List of the variables
Dependent variable Measurement Explanation
Livestock Output Euros This is the revenue of livestock farms. This variable takes the natural
logarithm of Livestock Output Independent variables
Capital Intensity Euros The amount of capital per unit of labour. This variable takes the natural logarithm of Capital Intensity
EU15 Dummy variable, either
the value of 1 or the value of 0
Value 1 = country is part of the
original EU15, value = 0 the country is not part of the original EU15
29 Methodology
This section includes a description about the quality of the data, the model specification and the descriptive statistics.
Quality of the data
The FADN data base provides panel data, the data is from a set of individuals (e.g. farms) over a number of years measured as an average economic indicator per member state. Different econometric tests are needed to test the quality of the data to make sure the results are interpreted correctly. The histograms of the variables do not show a normal distribution, leading to skewed distribution and violating the assumption of correct standard errors. Taking the log of the variables helps to eliminate the problem of an abnormal distribution. So, to normalize the data, the regression model will be estimated by using the natural log values of the variables, making it a log-log model. Both the dependent as the independent are
transformed by the natural logarithm. In order to use this model both variables must be greater than zero. All other variables have no negative values, therefore a log-log model can be used. A Hausman test is needed to identify whether a random or a fixed effect model is appropriate. The results of the test suggest that a random effect model is appropriate to use in the empirical analysis. By adapting a random effect model, which is the correct model according to available data, time invariant variables could be included. The variation in the random effect model is assumed to be random and uncorrelated with the independent variables included in the model. The Breusch-Pagan Lagrangian multiplier (LM) test for random effects also suggests that there is evidence to assume there is significant differences across countries and therefore it is appropriate to use random effects.
However, when testing for heteroskedasticity via the White test and autocorrelation via the Wooldridge test, results suggest that both where a problem in the model. The Pesaran cross-sectional dependence test is performed as well. Results suggest the same as the
Wooldridge test, hence there is enough evidence against the null hypotheses that residuals are not correlated. Therefore, there is cross-sectional dependence in the panel data. The model is estimated with robust standard errors to account for both problems.
The presence of multicollinearity has been investigated as well. The correlation matrix is included in appendix I. Results above 0,8 or below 0,8 are seen as multicollinearity
30 test was computed. Because the full model includes both a dummy variable and interaction variable which includes dummy variables as well, high VIF were the results of this approach. When looking at the other variable the results of the VIF test suggest that the VIF is not higher then 10 (informal rule of thumb according to Williams (2015)), thus collinearity is no problem in this thesis.
Finally, the Levin-Lin-Chu test is performed to check if the data is stationary. Results suggest that the null hypotheses of a unit root could be rejected, therefore the alternative hypotheses could be accepted and it can be concluded that the data is stationary.
Model description
The econometric model in this thesis tests the relationship between capital intensity and livestock farm performance for new and old member states.
For the model, the following equation is estimated:
(1) ln(Livestock Output) i,j = β0 + β1 ln(Capital Intensity)i,t + β2 (EU15)i,t + β3 (EU15 * Capital
Intensity)i,t + ui + eit
Where i, and t representing country and year and β1 – β3 are parameters to be estimated. The
error term eit capturing all events not including in the specification of the model and ui stands
for the country effects in the model. Livestock Output is the depended variable, measured in the natural logarithm. It represents the livestock output in agriculture of a particular country i
in year t .Capital Intensity is the first explanatory variables measured in natural logarithm. It is
the capital intensity per unit of labour. The expected relationship for this variable is positive. Based on the literature review, it is expected that the more capital available per unit of labour, the higher the livestock output will be. The second explanatory variable is the dummy
variable EU15 which either takes the value of 1 or 0. The expected sign is positive, being part of the EU15 results in a positive relation on livestock output. Lastly, β3 represents the
31 Descriptive statistics
The final sample consist of a total of 189 observations across 27 EU member states from the period 2007 until 2013. The benefit of using one data source is that the measurement is
consistent throughout the years, but as mentioned earlier the dataset is unbalanced because not each year the same farms are represented in the country average of a particular year. Table 3 shows the summary statistics used in the regression.
Table 3 Descriptive statistics
For Livestock Output the standard deviation is quite small, namely 1.090 with a minimum of 8.402, a maximum of 12.65 and an average of 10.46. For the variable Capital Intensity, the standard deviation was even smaller, namely 0.953 with a minimum of 6.247, a maximum of 10.35, and an average of 8.641. Lastly, the dummy variable is presented in the fifth row. This variable takes the value of either 1 (max) or zero (min).
(1) (2) (3) (4) (5)
VARIABLES N mean sd min max
Livestock Output 189 10.46 1.090 8.402 12.65
Capital Intensity 189 8.641 0.953 6.247 10.35
EU15 189 0.550 0.499 0 1
32 IV. Empirical results
This section presents the results from the regression with random effects in table 4. A robustness check with alternative independent variables is presented in table 5. Results
Table 4 shows the model estimated with random effects. It shows the level of significance and number of observations. However, it does not show the R-squared, as this is less useful when estimating a model with random effect compared to OLS regressions due to the different assumption. All the variables are in log form, except the dummy variables as described in the data specification. The first column represents the model as calculated in equation (1).
(1)
VARIABLES Livestock Output
Capital Intensity 0.233*** (0.0836) EU15 -4.585*** (0.979) EU15_Capital Intensity 0.576*** (0.127) Constant 8.052*** (0.666) Observations 189 Number of Countries 27
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 4 Estimated regression results with random effects
From table 4 it can be seen that all three explanatory variables are highly significant at a 1% significance level. When looking at the first row in the first column, Capital Intensity has the expected positive effect on Livestock Output. In the setting of the regression model, it can be interpreted as the percent change in y (Livestock Output), while x (Capital Intensity) increases by one percent. For Capital Intensity, it can be concluded that a one percent increase in the amount of capital per annual work unit in farms would yield a 0,23% increase in output of livestock and livestock products. Holding all other variables constant.
33 Unexpectedly, this coefficient has a negative effect on the output variable. This is not in line with the expectations nor the literature. So, being part of the EU15 and therefore classified as an old member state has a significantly negative effect on farm output. Holding all other variables constant. This negative results might be the result of using a random effect model in this research (no time invariant variables are omitted). Other econometric techniques could result into different output. Additional research on this matter is needed but not addressed in this thesis.
The interaction term EU15_Capital Intensity in the third row of table 4 reports a positive relation between livestock output, which is in line with the hypotheses stated in section II. This interaction terms indicates that old member states have a significantly higher amount of capital per labour unit which positively effects livestock output by approximately 0,58 % more, compared to new member states. To conclude, these results give enough
evidence to accept both hypotheses. There can be gains from an emission trade system for old and new member states because there are capital intensity differences in agriculture. And old member states have higher capital intensity in livestock farms compared to new member states. However, the extent to which there can be gains is hard to answer as the exact emission costs in agricultural production could not be derived from the dataset. This limits the
interpretation of the results as capital intensity only gives indirect evidence for energy differences across member states. A broader conclusion of the research will be given in the last chapter, together with limitations of the chosen approach. In the next section a robustness check is performed to stronger the results from this section. The robustness check is
performed with earlier mentioned alternative variables from the FADN database. Robustness check
In order to check the robustness of the results presented in this paper, a set of new independent variables are created to replace the original independent variables of the regression presented in table 4. The results of the new set of regression models with alternative independent variables are presented in table 5. All the continuous variables are measured in the natural logarithm. The results show similarities to the regression with the original dependent variable. When interpreting the results, there is still enough evidence to accept the hypotheses. The total fixed asset variable is significantly different from zero at a 1% significance level. The coefficient has the expected positive effect as well as the
35
(1) (2) (3)
VARIABLES Livestock Output Livestock Output Livestock Output
EU15 -6.136*** -1.155* -3.003***
(1.348) (0.689) (1.124)
Total Fixed Assets 0.329*** (0.0718) EU15_Total Fixed Assets 0.500***
(0.114)
Energy Costs 0.576***
(0.0441)
EU15_Energy Costs 0.164*
(0.0867)
Average Farm Capital 0.686***
(0.0561) EU15_Average Farm Capital 0.260*** (0.0980) Constant 6.139*** 5.219*** 1.995*** (0.761) (0.387) (0.641) Observations 189 189 189 Number of Countries 27 27 27
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
36 V. Conclusion and discussion
In the last section of the paper, conclusions are drawn and a policy discussion is given. This chapter ends with the limitations of the study and options for future research.
Conclusion
In this thesis, the method of De Cara and Jayet (2011) was used to provide a theoretical background to achieve cost-effectiveness for old and new member states in favor of a cap and trade system for agriculture. By showing graphically the implications of an emission
reduction scheme under the BSA, cost-effectiveness of emission reduction could be achieved (in theory) if such a system was applied to the agricultural sector. The MAC model underlines the need for flexibility, which is currently not given under the BSA. According to De Cara and Jayet (2011) a cap and trade system gives enough stimuli to emission reduction incentives of farm owners. More specifically, a panel data regression analysis is performed and results suggest that there is indeed a positive relationship between capital intensity and livestock output. Moreover, there is a significant higher positive relationship of the amount of capital per unit of labour on livestock output for old member states compared to new member states. Unfortunately, the effect of being part of the EU15 on livestock output was negative. This is not in line with the expectations. As explained in the result section, the random effect model allows for time-invariant variables to be estimated, rather than control for it as the fixed effect model does. Thus, other econometric tricks might derive different results and future research on this matter is needed. However, the significant and positive interaction variable provides enough evidence to accept both hypotheses. The relationship between capital intensity and livestock output is significantly higher in old member states than for new member states. It can be concluded that emission cost reduction possibilities can be achieved across EU
member states because there are capital differences for old and new member states. Therefore, there are gains from an emission trade system for EU member states from a cost-effective perspective. The extent to which there are gains from an emission trade system is hard to measure as the exact abatement costs for old and new member states are not available. The results and conclusions should be interpreted in the framework of assumptions of this study.
37 eastern and western European countries. They argue that new member states suffer from agricultural climate change policies due to their low adaptive capacity. The above literature suggests that the current policy structure is hugely debated under the effort sharing agreement. According to the literature there is not given enough flexibility to both old and new member states in adopting to agricultural climate response. Collective efforts are needed to meet the reduction targets in the most cost-effective manner via an emission trading system. This is in line with literature from Grosjean et al. (2016), who argue that there is an untapped potential for a market based system in agriculture and that such a system could play a critical role in incentives for cost-effective abatement in the European Union.
From a welfare perspective, it can be concluded from literature by Randon and Stavins (2015) that a potential cap and trade system could positively influence cost-effectiveness across member states. Research from Randon and Stavins (2015) state that if both countries accept the allowances and participate in international trade, it creates benefits for both
countries and will lead to welfare gains. The authors argue that the most important reason for linking systems is the increase in cost-effectiveness that results from the allocation of
abatement costs between systems with different MAC curves. Perez-Dominguez et al. (2009) are also in favor of a more flexible emission abatement possibilities, such as a cap and trade. According to the authors this induces economic profits and welfare gains for society.
In the end, it is the trade-off between the distortion effects of the introduction of such a system and efficiency gains. However, more research is needed to investigate if the efficiency gains outweigh the distortion effects. A more detailed analysis is required to shine a new light on the possible burden reduction options for the non-ETS sectors via an European emission trade system in the context of the EU BSA.
Policy discussion
38 of the EU is needed to successfully launch such a system. It is important that the credibility and transparency of the DO projects is insured. With adequate regulatory requirements transaction costs will be limited. It is desirable that these projects will eventually create entrepreneurship and climate awareness in the agricultural sector in regard to emission reductions and energy intensity. However, possibilities to ensure such kind of incentives are not discussed in this paper, as this is unfortunately beyond its scope. Likewise, potential project methodologies for agriculture are not covered. In the future, more research is needed to investigate the further implementation of the project-based crediting scheme in agriculture. A more detailed analysis of the difference between linking the non-ETS with ETS or the set-up of a sectorial non-ETS cap and trade system is needed, but not addressed in this study. Limitations
There are some limitations to the research presented in this paper which are outlined in this section. The main restrictions of this study relate to the data. First of all, there is a limited number of observations due to the nature of the data. Only country averages where available due to privacy sensitivity of the data used in the FADN database. This limits both the quality and the estimation of the coefficients of the model. A more specific estimation of farm performance across member states could have been made if data were available per company (per farm in this case), extending the number of observations and therefore the quality of the model. Another limitation of the research is the measurement and definition of both input and output variables used in the regression analysis. The precision in estimating might be limited as a result of available measures in the dataset. If more specific variables where available, especially more information about emission use in the input variables and as part of the total output, the potential emission reduction between member states could be estimated more precisely. Consequently, to the nature of the data, both autocorrelation and heteroskedasticity where a problem, which is accounted for by using robust standard errors. The regression results could still be debatable due to the presence of both autocorrelation and
heteroskedasticity in the model. An assumption is made on the linear relationship between dependent and independent variables due to the way estimated coefficients are put in the model. However, it might not be the case for the model to be linear, which makes the
39 Future research
For the future, more research can be done by comparing Europe with other continents of the
world on adopting emission reductions in agriculture, such as North-America or Australia. By extending the research across the borders of the EU, it might give different insights on the matter. Another option is investigating other non–ETS sectors besides agriculture on the need for collective efforts to reduce emissions between member states. For example, the transport or build sector could be investigated. Furthermore, different econometric techniques could be used to improve and indicate the differences in input and output measures in the agricultural sector between new and old member states. Other determinants of farm inputs on farm output could be studied to give a broader scope to the research.
Acknowledgement
40 References
Aldy, J.E., and Stavins, R.N. 2012. The Promise and Problems of Pricing Carbon: Theory and Experience. Journal of Environment & Development 21 (2), 152 – 180.
Betz, R., Rogge, K., and Schön, M. 2006. Domestic Offset Projects: Limited opportunities in Germany but potential for others? Energy & Environment 17 (4), 569 – 582.
Bouma, J., Varallyay, G., and Batjes, N.H. 1998. Principal land use changes anticipated in Europe. Agricultural. Ecosystem Environment 67, 103 - 119.
Cole, M., and Elliott, R.J.R. 2005. FDI and the Capital Intensity of ‘‘Dirty’’ Sectors: A Missing Piece of the Pollution Haven Puzzle. Review of Development Economics 9 (4), 530-548.
De Cara, S., Houzé, M., and Jayet, P.-A. 2005. Methane and nitrous oxide emissions from agriculture in the EU: a spatial assessment of sources and abatement costs. Environmental and Resource Economics 32 (4), 551–583.
De Cara, S., and Jayet A. 2011. Marginal abatement costs of greenhouse gas emissions from European agriculture, cost-effectiveness, and the EU non-ETS burden sharing agreement. Ecological Economics 70 (9), 1680–1690.
Curtiss, J. 2000. Technical Efficiency and Competitiveness of the Czech Agricultural Sector in late transition– The Case of Crop Production. Paper presented at the KATO Symposium Berlin, Germany, November 2–4. [accessed 23-01-2017]
https://www.academia.edu/275094/Technical_Efficiency_and_Competitiveness_of_the_Czec h_Agricultural_Sector_In_Late_Transition._The_Case_of_Crop_Production
Davidova, S., Gorton, M., Iraizoz, B., and Ratinger, T. 2003.Variations in farm performance in transitional economies: Evidence from the Czech Republic. Journal of Agricultural Economics 53, 173–361.
Dos Santos, M.J.P.L. 2013. Segmenting farms in the European Union. Agricultural Economics 59 (2), 47-57.
