Financial Economics 
Is the EU ETS effective? 
Examining the impact of the EU ETS on methane emissions 
 
Bachelor Thesis 15hp 
         Authors: 
         Carl Otto Lindsjö 
         William Karlander 
         Tutor: 
         Jon Williamson 
         Mentor: 
         Suad Januzzi STX Group 
         Spring term 2023 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
ii 
 
 
 
 
Abstract 
The European Union Emission Trading System (EU ETS) is the world's largest trading system 
for emissions. It has a crucial role to play in meeting the goals of the Paris Agreement. 
However, there has been debate over its effectiveness. This thesis examines how the EU ETS 
impacts methane emissions. The study uses Swedish emission data from 2005 to 2021 and 
employs a fixed effect model with two exogenous and two endogenous variables. While the 
test produced significant results for carbon dioxide emissions, it could not do so for methane 
emissions. The thesis concludes that EUA prices have a negative effect on carbon dioxide 
emissions, while electricity prices have a positive effect. The impact on methane emissions 
could not be determined. Previous research is divided, suggesting that financial solutions like 
a cap and trade system can have varying effects depending on the sector and perspective. A 
new model with interaction variables, more control variables, or a different data set could yield 
significant results. 
 
Acknowledgement 
This endeavour would not have been possible without the assistance and supervision of PhD 
Jon Williamson, to whom we would like to express our utmost gratitude. Williamson provided 
the necessary feedback to help us structure and complete this thesis. We were lucky to have 
had direct contact with Ravigné for permission to cite his working paper, which proved 
invaluable. We also thank our fellow students for their feedback and support while writing this 
thesis. Lastly, we would like to thank Suad Januzzi for providing the idea of this thesis and 
valuable insights into the matter.  
  
  
  
  
  
iii 
 
 
 
 
 
 Table of Contents 
1. Introduction 1 
1.1 Problem Description and Purpose 4 
1.2 Research question 5 
2. Foundation 6 
2.3 Law of demand 6 
2.4 Rebound Effect 7 
2.5 Backfire Effect 7 
2.6 Spillover effect 8 
2.7 The Effect of Financial Solutions on Emissions 8 
3. Data 11 
4. Research design 16 
4.1 Model specification 18 
4.2 Standard errors 19 
5. Result 20 
6. Discussion 22 
7. Conclusion 27 
Reference list 29 
 
 
 
  
iv 
 
1. Introduction
It has not escaped anyone that global warming is a problem. Global warming is characterised
by an increase in Earth's average surface temperature caused by human activities, such as
burning fossil fuels that emit greenhouse gases into the atmosphere. Global warming poses a
significant risk as temperature increases raise the risk of heatwaves, diseases and droughts.
Scientists agree that global warming will continue to increase if we do not decrease the
emission of greenhouse gases (NASA, 2011). Carbon dioxide is the most common
greenhouse gas emitted (Worldbank, 2020a). It is emitted from a wide range of sectors, and
there are systems in place to reduce its emissions, such as carbon taxes and the EU ETS.
However, other gases, such as the second most emitted gas, methane (Worldbank, 2020b),
contribute to global warming without such systems. When comparing the effects of these
greenhouse gases, their respective greenhouse warming potential (GWP100) is used. The
GWP100 is calculated as the 100-year impact on global warming, where carbon dioxide is set
to 1, and one ton of methane equals 25 (Eurostat, 2017).
The Paris Agreement, adopted in 2015, provides clear guidelines based on GWP100 to limit
global warming to a maximum of 2°C by 2050 (United Nations, 2023a, n.d). However, a
shortcoming of the GWP100 is that it fails to account for the short-lived effects of methane.
The effects of methane emissions are greater during the first 20 years, with a warming effect
of up to 80 times greater than carbon dioxide (UNEP, 2022). Acknowledging the short-term
effect shows that it is crucial to reduce methane emissions to meet the Paris Agreement's
requirements. To reduce methane emissions, current procedures must transition into new,
more sustainable operations (Formas, 2023). In order to transition into new, more sustainable
operations, stakeholders need to be held accountable for their contributions to global
warming. Holding stakeholders accountable is usually done through financial systems, such
as taxes, or cap and trade systems, such as the EU ETS. However, few systems are put in
place to reduce methane emissions. Therefore, it is essential to investigate whether new
systems are needed or if carbon-based systems such as the EU ETS are sufficient.
1
Exploring current financial solutions can help determine if new solutions are needed. A
carbon tax is one of the more common financial solutions used. Carbon taxes set a price on
carbon dioxide emissions that polluters must pay (World Bank, n.d.).
Another financial solution proposed to combat global warming and reduce greenhouse gas
emissions is the European Union Emission Trading System (EU ETS). The EU ETS is a
cap-and-trade system where greenhouse gas emissions have a cap that is reduced each year,
which should reduce emissions if it has the intended effects. The EU ETS was adopted as a
policy instrument to meet the targets of the Kyoto Protocol established in 1997 (European
Commission, n.d.a). The idea of the EU ETS is to force reductions in emissions from the
energy sector, the manufacturing industry and aircraft operations. It covers around 40% of the
EU's total greenhouse gas emissions and sets a cap for mainly carbon dioxide emissions. The
emission allowances used in the EU ETS are called European Union Allowances (EUA).
Entities participating in the EU ETS must hold enough EUAs to cover their emissions at the
end of each year; otherwise, fines are imposed. A company's excess EUAs can be sold to
other companies within the EU ETS (European Commission, n.d.b), rewarding companies for
reducing their emissions.
There are plans for the EU ETS to include more sectors, such as the shipping industry, by
2024, and other greenhouse gases, such as methane emissions, for the shipping sector in 2026
(European Commission, n.d.c). This will likely be the first cap and trade-based system to
reduce methane. There are other types of regulatory policies that strive to reduce methane
emissions, but seemingly no financial solutions yet, such as a system like the EU ETS for
carbon dioxide or voluntary methane credits allowances. A part of this gap will be addressed
by including methane gas in the EU ETS, but other solutions might be needed.
One can look at the quota between carbon dioxide and methane emissions to determine if
new systems are needed to reduce methane emissions. The quota will provide insight into
whether both gases are being reduced at the same rate, as a change in quota would mean that
one of the gases reduces or increases at a different rate. If a carbon dioxide-based system
reduces carbon dioxide, then ceteris paribus, the quota between methane and carbon dioxide
emissions should increase. If, on the other hand, the quota holds constant, then it could
indicate that the system also affects methane emissions. To illustrate this, the quota between
methane and carbon dioxide emissions in the European Union (EU) is calculated from the
data obtained by the World Bank (2020a, 2020b). The data is presented in CO2 equivalents
2
calculated by the GWP100 between the years 1990 to 2019. In 2019 carbon dioxide
accounted for 80.1%, and methane accounted for 10.8% of total greenhouse gas emissions in
the EU. Figure 1 shows the quota for methane emissions and carbon dioxide between 1990 to
2019 within the EU.
Figure 1
Notes: The quota between the total amount of methane emissions and carbon dioxide emissions within the EU
from the starting year 1990 until 2019
As shown in Figure 1, the quota between methane and carbon dioxide emissions decreased
between 1990 to 2005, indicating that the emissions changed at different rates. The trend
shifted in 2005, indicating that the reduction rate changed between the gases. This could be
because the EU ETS was implemented in 2005, causing reductions in carbon dioxide
emissions. Figure 1 indicates that the EU ETS only affects carbon dioxide, but this is a
simplification as many factors can affect the emission levels. Thus no conclusion can be
drawn. Isolating the effect of the EU ETS will answer if the system affects methane
emissions in any way.
3
1.1 Problem Description and Purpose
As mentioned in the introduction, methane emissions significantly contribute to global
warming. Reducing methane emissions is, therefore, crucial to achieving the goals set out in
the Paris Agreement. Although there are established solutions for reducing carbon dioxide
emissions, such as carbon taxes, the EU ETS and voluntary carbon credits, it is unclear if
these solutions affect methane. It is also unclear if methane emissions would reduce if
methane was included in the EU ETS, which is interesting to examine since it will be
included in the shipping sector by 2026. Previous research has indicated varying impacts
regarding how a carbon tax or a methane-based cap and trade system could affect other
greenhouse gas emissions.
According to research by Key and Tallard (2011), implementing a carbon tax and a
methane-based cap and trade system could reduce methane emissions. However, there exists
conflicting research. The backfire effect, by Ravigné and Naudad (2023), demonstrates that
natural gas emissions can increase when carbon emissions are reduced. Additionally, Goulder
and Schein (2013) have highlighted that introducing another greenhouse gas into a cap and
trade system may be ineffective, as pollution may shift to other areas. The ineffectiveness of
introducing further regulation also holds for emission policies if a cap and trade system is
already in place. Research by Key and Tallard (2011) suggests that a carbon tax and a
methane-based cap and trade system could reduce methane emissions. The research by Key
and Tallard (2011) contrasts the backfire effect mentioned by Ravigné and Naudad (2023).
The backfire effect demonstrates that when a reduction in carbon emissions is made, there
could be increases in natural gas emissions (methane). Goulder and Schein (2013) have
shown that introducing another greenhouse gas in a cap and trade system may be ineffective,
as pollution may increase elsewhere, and that it is also ineffective to introduce regulation
when a cap and trade system is already in place. Testing if the EU ETS already affects
methane emissions could help determine if there is a rebound, backfire or spillover effect and
also give policymakers more information when creating new financial solutions to reduce
methane emissions.
4
1.2 Research question
In order to fulfil the purpose of this thesis, the following research question will have to be
answered.
What impact does the price of emission allowances in the EU ETS system have on methane
emissions?
To answer the research question, the following hypothesis is tested.
Ho: The EUA price does not have an effect on methane emissions
Ha: The EUA price has an effect on methane emissions.
5
2. Foundation
2.1 Exogenous price settings
Before further examining how emissions can be reduced through financial solutions, it is
essential to explore the underlying theories that these solutions are based on. One crucial
concept in reducing emissions through measures like emission taxes or, to some extent, the
EU ETS is the concept of exogenous price settings. Exogenous price setting refers to price
floors and limits that make producers shift their production levels to a new optimum.
Exogenous price settings occur when prices are determined externally to the market by
factors such as regulations or policies (Perloff, 2021). Exogenous emission price settings are
usually in the form of a limit, where the carbon tax is the most common. The EU ETS uses an
exogenic price setting as a cap but could be considered a more nuanced approach due to the
market deciding the prices.
2.2 Neoclassical theory
Neoclassical theory is another fundamental theory that financial policies rely on and sheds
light on why emissions are not naturally reduced without intervention from external actors.
"The Neoclassical model assumes that firms maximise their profits from producing and
selling goods and services" (Goodwin et al., 2022). If any financial solution is to be effective,
firms need to follow the principles of profit maximising. The assumption also means that
firms are cost-effective and strive to find the most efficient way to reduce costs. Since
emissions are costly, profit maximising might mean the producers ignore their emissions and
let another market participant stand for that cost. For the case of evaluating the result of this
thesis, one central assumption that will be utilised is that firms act rationally and are
profit-maximising.
2.3 Law of demand
A second classic economic concept that may be used as support for a financial solution is the
law of demand. The law of demand states that if all other things are equal, the quantity of a
good or service demanded by consumers is affected by price in a linear relationship where
6
demand increases as prices decrease (Hayes, 2022). Suppose a good or service is elastic,
meaning its demand changes significantly in response to price changes. In that case, it is
likely that the consumption of that good or service will decrease if emission prices increase
due to regulatory policies such as carbon taxes (Hayes, 2022). This principle can be applied
to any emission, including methane, as Key and Tallard (2011) described for the agriculture
sector. Consumption of emission-intensive products will decrease because consumers
substitute them for cheaper, less emission-intensive goods or services. Key and Tallard (2011)
showed that methane emissions have a price elasticity of 0.02 at a carbon dioxide price of
USD 20/tCO2-eq. The positive price elasticity means that methane emissions within the
agriculture sector are shown to decrease when the price of carbon dioxide emissions
increases. Therefore, the law of demand can explain how price fluctuations affect methane
emissions.
2.4 Rebound Effect
Previous research states three effects that describe how solutions affect emissions. The
rebound effect is a concept where environmental improvements may increase consumption.
Increased reductions within consumption areas, such as electricity consumption, lower the
usage cost, leading to increased electricity consumption (Umweltbundesamt, 2019). The
rebound effect is important when discussing emissions, especially concerning the EU ETS,
which does not capture global emissions. If energy efficiency improvements lead to increased
consumption, the emission reductions achieved may result in little or no net reduction. The
rebound effect shows a limitation of the study. Even if there is an effect, it could be the case
that the emissions were rebounded to another country outside of the EU.
2.5 Backfire Effect
The second effect described by previous research is the backfire effect. The backfire effect is
a theory describing the phenomenon where a decrease in carbon dioxide emissions may
backfire and lead to increased methane emissions (Ravigné & Nadaud, 2023). Ravigné and
Nadaud give an example of a carbon tax that did reduce carbon dioxide emissions. Reducing
carbon dioxide emissions may render it successful, but Ravigné and Nadaud show that
natural gas consumption increased after introducing the tax. The backfire effect could be an
7
issue for EU ETS, because the intended carbon reductions lead to increased methane
emissions. Suppose that the EU ETS has a positive effect on methane emissions. Then the
backfire effect could explain this positive effect.
2.6 Spillover effect
The third effect described by previous research is the spillover effect. The spillover effect is
the impact that events or actions in one country or economy can have on other countries or
economies. These effects can be positive or negative and are caused by many things which
have become more prevalent in the more interdependent world (The United Nations'
Sustainable Development Solutions Network, 2022). Siping et al. (2021) describe that carbon
dioxide positively correlates with other emissions. They consider it a risk because carbon
dioxide emissions are rising in China. Therefore other emissions may too because of the
spillover effect. However, this effect could also explain why decreased carbon dioxide
emissions reduce methane emissions. Therefore, The concept helps explain the potential
effects of EU ETS on methane.
2.7 The Effect of Financial Solutions on Emissions
The three effects described previously help explain the possible effects of the EU ETS.
However, further examining previous research is needed to gain a deeper understanding of
the potential effects of the EU ETS. In Carbon taxes versus cap and trade: A critical review
(2013), Goulder and Schein look into the performance of two financial systems, a carbon tax
compared to a cap and trade system. They show that with an established emission price, in
this case, carbon dioxide, both taxes and cap and trade systems incentivise firms to change
their production processes to reduce emissions. These policies also affect consumers through
the law of demand, encouraging changes in their consumption to less carbon-intensive
emission goods. The way this change is made differs from the policy used to increase the
price. With a tax, the emissions price is fixed according to the tax rate (exogenous price
setting), whereas in a cap and trade, the price is more indirect and set by the market.
Regardless of the policy used, the firms affected by them face the same price of emissions,
8
which, if firms are cost-minimising according to the neoclassical theory, then emissions will
be reduced to the margin where the cost of emissions is equal to the emission price.
Goulder and Schein (2013) also compare the relative advantages and disadvantages of the
two financial solutions. One relative disadvantage mentioned for a cap and trade system is
that the amount of allowances is inelastic, which means that price changes of emissions due
to demand shifts can produce volatility (Nordhaus, 2007). The critical takeaway from price
volatility is that uncertainty about emissions prices under cap and trade systems constrains
the firm's ability to respond to any changes in climate policies.
Another topic discussed in their thesis is whether introducing a new limit on any greenhouse
gas has an effect. For the cap and trade system, they argued that the effect of introducing a
new greenhouse gas to the trading system or a limit to emissions might be ineffective
(Goulder & Schein, 2013). They demonstrated the ineffectiveness of introducing another
limit through an example of the United Kingdom's (UK's) electricity production after
introducing a carbon tax. The UK energy producers had to pay an additional fee to the UK
government, apart from the EU ETS allowance price. The amount of emissions from UK
energy production was reduced. However, the overall pollution from energy production in
Europe would be shown to rebound, leading to increases that would equal the diminish from
the UK. The rebound effect shows that the pollution could increase if the production is in
countries with more lax policies or if they are excluded from the EU ETS.
The rebound effect is important since a significant flaw with the cap and trade system is
shown. The effect of introducing another emission-limiting policy will be offset by the effect
of any cap and trade system such as the EU ETS. The policies merely relocate the emissions
to areas outside the country rather than reduce them. The effect of the policies in a country
also depends on the influence it can exert over the cap and trade system imposed on it. A
sufficiently large country influences the cap and trade system, and they can therefore affect
the system in their favour.
The rebound effect showed that existing policies may have unintended effects on new
solutions. It is, therefore, essential to investigate the existing policies and financial solutions
targeting methane emissions. This includes analysing any interactions or possible causes of
the result, as highlighted by Goulder and Schein (2013). In this context, Key and Tallard
(2011) examine whether a carbon tax may reduce methane emissions in the agriculture sector
and if a methane-based cap and trade system reduces methane emissions.
9
Key and Tallard (2011) use a recursive-dynamic partial equilibrium model (this will not be
explained in this thesis, but it may be interesting to know what model was used) of global
agriculture to investigate whether a carbon tax and a cap and trade system similar to the EU
ETS based on average national methane could be effective. They base their study on a
methane-based sectoral cap and trade system, where allowances flow between countries, not
companies. Therefore, the companies do not have any incentives to reduce their emissions.
Key and Tallard (2011) demonstrate that if Annex 1 countries imposed a carbon tax in the
agriculture sector, methane emissions would decrease by 3.9%. Therefore, a carbon tax has a
negative effect on methane, which could mean that the EU ETS also has a negative effect on
methane.
Their thesis is built upon a linear relationship between production (in the agriculture sector)
and methane emissions. The study shows that methane emissions are reduced when a carbon
tax is used. The effect depends on if the carbon tax is global or just on industrialised countries
(Annex 1). When taxes are limited to Annex 1 countries, Key and Tallard's model shows that
two-thirds of the reductions will rebound through increased production in other countries. In
either case, introducing a carbon tax or a cap and trade system based on methane emissions
for each specific country could reduce methane emissions according to their model. An
important highlight of their study is that the effectiveness of the policies may vary depending
on what strategy is implemented. Suppose every country uses a similar strategy, such as a tax.
In that case, the effect of the cap and trade system will remain constant regardless of how
high the cap is. This also means that the effectiveness of a cap and trade on methane may
depend on the respective countries' methods to reach these targets. The fact that cap and trade
systems are shown to have different effects in different countries makes it challenging to
create a single model that can explain how the EU ETS affects emissions in all countries.
However, Key and Tallard (2011) describe that Annex 1 countries are similar and seem to use
the same solutions to reduce emissions. Therefore, a cap and trade system would have similar
effects within Annex 1 countries.
10
3. Data
In order to create and test a model of the effectiveness of the EU ETS, it is essential to collect
reliable data on emissions, free from significant inaccuracies. The model should also produce
results that can be applied to other populations. Key and Tallard (2011) suggested that the
effects of a cap and trade system should be comparable for all Annex 1 countries. As Sweden
is an Annex 1 country, its data can be used to generalise the effects of the EU ETS on other
Annex 1 countries. The Swedish Environmental Protection Agency (2022) has assessed
Sweden's data according to Swedish and EU laws, making it more reliable.
The data used for this thesis is publicly available. Thus we deem the ethical risks of the data
to be non-existent. The rest of the thesis should be in accordance with research ethics as
described by 4 § Lagen om ansvar för god forskningssed och prövning av oredlighet i
forskning (SFS 2019:1150).
The data used in this thesis is based on carbon dioxide and methane emissions. The data on
carbon dioxide is gathered from the Swedish Environmental Protection Agency (2022) and is
categorised by sector. The data covers seven sectors, collectively accounting for 94% of
Sweden's total emissions. These sectors include industry (27%), domestic transport (28%),
foreign transport (13%), agriculture (10%), energy (8%), machines (5%), and waste (3%).
The data is given as carbon dioxide equivalents for each sector annually between 2005 and
2021.
A limitation of the data from the Swedish Environmental Protection Agency (2022) is that it
only includes total carbon dioxide equivalents for different sectors and does not provide
information on methane emissions. Therefore, methane levels need to be estimated for each
sector. One way is to use quotas for each sector. However, it is difficult to determine accurate
quotas as there is a lack of data on methane emissions for each sector. Therefore, data from
the World Bank (2020b) on Sweden's methane emissions levels are used between 2005 to
2019. The national quota between methane and carbon dioxide equivalents is calculated
annually between 2005 and 2019. There is no data on methane emissions after 2019.
Therefore, the quota between methane and carbon dioxide equivalents is assumed to remain
constant between 2019 and 2021. These quotas are then applied to all sectors in Sweden. This
may affect the reliability of the model if sector-specific effects are being tested, as quotas are
likely to differ between sectors.
11
The variable of interest in the model is the EUA price. It is obtained from the World Bank
(2023). The EUA price is given as yearly averages from 2005 to 2021, presented in US
dollars.
Lastly, Electricity price is used as a control variable. Further explanation as to why it is
chosen as a control variable is explained under Chapter 4, Research Design. Data on the
electricity price is gathered from Nordpool Group (2023). The data represents Sweden's
yearly average household electricity prices between 2005 and 2021.
In 2011 Sweden underwent changes in the electricity market. 4 electricity regions were
created, denoted as SE1, SE2, SE3, and SE4 (Energimarknadsinspektionen, n.d.). This means
that the data of the prices prior to 2011 are for the whole of Sweden, whereas after 2011, it is
divided. An average of the electricity prices was created for the data after 2011 in order to
structure the data. For 2011 there is no data. Therefore, an average of the prices in 2010 and
2012 was calculated.
All variables used in the model are described in Table 1.
12
Table 1 Variable description
Variable Description
CO2 Annual Carbon dioxide emissions (Mt)
CH4 Annual methane emissions, (Mt) carbon dioxide
equivalents.
EUA_price Average yearly price for one EU ETS allowance in
US dollars.
Electricity_price Average household electricity price in Sweden
(EUR/MWh)
lnCO2 Log of CO2
lnCH4 Log of CH4
The described variables' average characteristics are presented in Table 2. The table displays
the observations, mean, standard deviation, and minimum and maximum values. The dataset
contains emission levels during 17 years divided by seven sectors, summarising 119
observations. The dataset also contains the EUA prices and electricity prices.
13
Table 2 Descriptive statistics
Variable Obs Mean Std. Dev. Min Max
CO2 119 10.373 7.117 .99 24.14
CH4 119 1.393 .954 .135 3.153
EUA_price 17 17.279 12.42 1.26 49.78
Electricity_price 17 37.867 11.064 18.957 57.89
Before creating a model, it is important to investigate if the data suffer from multicollinearity.
This is done by calculating the correlation between the variables. Table 3 presents the
correlation between the exogenous and endogenous variables, specifically the correlation
between EUA_price and Electricity_price. According to the table, the correlation coefficient
is calculated as 0.684, indicating that the dataset does not exhibit perfect multicollinearity
(Stock & Watson, 2020). If the correlation between EUA_prices and Electricity_price
exceeds 0.8, it may pose challenges in measuring the individual effects of each variable.
Table 3 Correlation matrix
Variables (1) (2) (3) (4)
(1) CO2 1.000
(2) CH4 0.999 1.000
(3) EUA_price 0.001 0.001 1.000
(4) Electricity_price 0.043 0.026 0.684 1.000
Another way to show the correlation between the variables and their underlying trends is
displayed graphically in Figure 2. The electricity and EUA prices display high variance
compared to the emission levels. EUA and electricity prices seem to correlate but are not
perfectly collinear.
14
Figure 2
Notes: Graph of EUA prices, electricity prices, carbon dioxide and methane emission. The Y
axis represents the dollar price for EUA prices, SEK for electricity and million tonnes for
methane and carbon dioxide.
15
4. Research design
With the data specified, the research design could be conducted. This analysis aims to isolate
and analyse the EUA prices' effect on methane emissions. There are many challenges, as
there are a multitude of factors that can affect emissions. These factors need to be considered
to isolate the effect of the EUA price (Löfgren et al., 2013).
Previous research by Löfgren et al. (2013) used a difference-in-difference OLS estimator
when analysing the effect that the implementation of the EU ETS had on green research and
development. Using a difference-in-difference estimator was unsuitable for this thesis
because difference-in-difference estimators are used to analyse the effects of policy changes
and not price fluctuations.
It can be challenging to isolate the effect of the EU ETS by controlling for other
environmental policies because no data is available to match with the sector-level data
(Löfgren et al., 2013). Environmental policies are assumed to be equal within sectors but
differ between sectors and can therefore be controlled using a fixed effect model (Löfgren et
al., 2013). Thus, the fixed effect model was used to control the sector-specific policies.
An assumption for the data was made that there is a linear relationship between the prices of
carbon credits and emission levels. OLS regression is therefore preferred when analysing the
effects of price changes, as the OLS estimator captures their linear relationship. OLS
regression is also the preferred choice for economists and statisticians. (Stock & Watson
2020, pp. 151-152).
OLS regression is valid if the coefficients are consistent and unbiased. Sources of bias arise
when the regressors are correlated with the error term. Therefore, the OLS regression model
is successful if the statistical inference can be internally and externally valid (Stock &
Watson, 2020, pp. 330-331). The model is externally valid if the statistical inference is valid
for the population studied (Sweden), and that it can be generalised to other populations, such
as Europe (Stock & Watson, 2020, p. 331).
For the model to be internally valid, the regressors need to be unbiased and consistent (Stock
& Watson, 2020, p. 333). This means that EUA_price cannot correlate with the error term,
which can happen if relevant variables are omitted. To avoid this, the model must include
variables that can explain emission level variation and correlate with EUA_price. A tradeoff
16
is made when including explanatory variables, as including relevant control variables does
reduce the risk of OV bias, but it also increases the variance of the model. (Stock & Watson,
2020, pp. 334-335)
Löfgren et al. (2013) included fuel intensity as an instrumental variable for fuel price. They
argued that the price of fuel could dictate if companies would adopt new emission-reduction
technologies. Since our model does not test how the EU ETS changes investment behaviour
but tests if and how the price of emission allowances affects emissions, a different control
variable was needed. One variable that was more interesting than the fuel price for our model
was the electricity price. As Löfgren et al. (2013) mentioned, "There is a risk that fuel prices
suffer from endogeneity problems because investment in abatement technology will decrease
the use of fossil fuels." Swedish electricity prices are based on European demand, not solely
Swedish demand (Vattenfall, 2021). Therefore, the Swedish emissions do not explain the
variation in the Swedish electricity prices. Thus, Swedish Electricity prices should not have
the same endogeneity problem as fuel prices. Therefore, the electricity price is a valid control
variable as it can explain emissions and correlate with the EUA price. It may, however, be
problematic to assume that Swedish electricity prices are exogenous because Sweden is a part
of Europe. Therefore the Swedish emissions do, in fact, affect Swedish electricity prices to
some degree, albeit the effect is small compared to the effect of all of Europe's emissions.
Suppose we assume that the energy sector aims to maximise its profit as neoclassical theory
describes. In that case, the energy sector will increase its output when facing high household
electricity prices as this would increase its profits. Increased output increases emissions,
meaning that electricity prices affect emissions. Therefore we argued that the Swedish
household electricity price is a relevant control variable.
Löfgren et al. (2013) used fuel intensity as an instrument for fuel price to limit
multicollinearity. As shown in Table 3 and Figure 4, electricity price correlates with EUA
price. However, it is not perfectly collinear, meaning the variables do not suffer from perfect
multicollinearity but rather an imperfect multicollinearity (Stock & Watson, 2020, pp.
230–231). Imperfect multicollinearity does not threaten the OLS estimator because it does
not bias the coefficient estimates. However, the Standard error of the coefficient of EUA
price may increase (Stock & Watson, 2020, pp. 230–231).
The OLS estimator can fail to estimate the true effect of the independent variable if the
exogeneity assumption fails. The exogeneity assumption can fail if the variation of the
17
independent variable (EUA_price) can be explained by the dependent variable (emissions)
(Stock & Watson, 2020, pp. 427-429). The model would break the exogeneity assumption if
it were based on the entire EU because the level of emissions decides the demand for carbon
credits, which in turn explains the price of the credits, thereby violating the exogeneity
assumption (Stock & Watson, 2020, p. 427). To limit this endogeneity risk, we limited the
study to Sweden because the Swedish demand for carbon credits is small compared to the
demand of the whole European region. Therefore, we assumed that the independent variable
EUA_price is exogenous because emission levels in Sweden do not explain the demand for
carbon credits. An issue with this reasoning is that the underlying macro shocks determining
demand in the EU and Sweden may be correlated. Swedish demand predicts the Swedish
macro shocks, which could predict the EU macro shocks. The EU shocks can predict the
emissions within the EU, which is one factor that decides the EUA price. Thus, the Swedish
demand could predict the EUA price.
4.1 Model specification
Given the arguments mentioned, we specified the models as follows:
𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛𝑠  =  β 𝐸𝐴𝑈 𝑝𝑟𝑖𝑐𝑒  + β 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦 𝑝𝑟𝑖𝑐𝑒 + α + 𝑉 (1)
𝑖,𝑡 1 − 𝑡 2 𝑡 𝑖 𝑖,𝑡
Where t is the time in years, i is the sector, V is the Error term and α is the fixed effect for
each sector.
The sectors have different fixed effects denoted A, which we assumed do not change over
time. We, therefore, wrote 𝐴 = 𝐴. We assumed that A is a parameter and not a random
𝑡
variable. We, therefore, replaced A with αi.
We defined a dummy variable D(i)j,t for each sector which takes the value 1 if j=i or 0 if j≠i
The model could therefore be written as follows:
𝑛
𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛𝑠  =  β 𝐸𝐴𝑈 𝑝𝑟𝑖𝑐𝑒  + β 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦 𝑝𝑟𝑖𝑐𝑒 + ∑ α 𝐷(𝑖) + 𝑉 (2)
𝑖,𝑡 1 − 𝑡 2 𝑡 𝑗 𝑗,𝑡 𝑖,𝑡𝑗=1
18
We write:
𝑛
α = ∑ α 𝐷(𝑖)     (3)
𝑖 𝑗 𝑗,𝑡
𝑗=1
Plucking equation (3) in equation (2) gives our model (1)
𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛𝑠  =  β 𝐸𝐴𝑈 𝑝𝑟𝑖𝑐𝑒  + β 𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑖𝑡𝑦 𝑝𝑟𝑖𝑐𝑒 + α + 𝑉 (1)
𝑖,𝑡 1 − 𝑡 2 𝑡 𝑖 𝑖,𝑡
Where t is the time in years, i is the sector, V is the Error term and α is the fixed effect for
each sector.
A possible drawback of the fixed effect model is that we do not know the differences between
the different sectors. These differences affect emissions, but as we do not know what they are,
we will likely underestimate or falsely estimate the effect that EUA_price has on emissions
(Stock & Watson, 2020). The case of an underestimating or falsely estimating model will be
discussed further under the discussion section.
4.2 Standard errors
Another thing to keep in mind with the model is that when conducting this model in Stata,
Stata will assume that all data points are independent of one another. This would rule out that
a decrease in emissions today would affect emissions tomorrow (positive auto-correlation).
Which likely is not true. The data table is indexed by a sector identifier (i) and a time period
identifier (t).
Assuming that all data points are independent of one another rules out that a decrease in
emissions today would affect emissions tomorrow (positive auto-correlation), this is not very
plausible. It is more plausible that data points within sectors are dependent on one another but
independent between sectors. Therefore we assumed cross sectional independence, i.e that
Vi,t and Vi´,t are statistically independent for 𝑖 ≠ 𝑖' , and not restrict the temporal𝑡
dependence within a sector, thereby leaving cov(Vi,t , Vi,t’) unrestricted. Therefore, we
clustered the standard errors on a sector level which will increase the standard errors and
p-values of the model as the model now allows for heteroskedasticity and autocorrelation
within a sector. (Stock and Watson 2020, 375-376).
19
5. Result
Two models were created to examine the EUA price's effect on methane emissions and
carbon dioxide. Model 1 examines the effect of the EUA price on methane emissions, and
Model 2 examines the effect of the EUA price on carbon dioxide. An overview of the results
is shown in Table 4.
Table 4 Fixed effect regressions on the log of methane emissions (1) and log of carbon
dioxide emissions (2), with clustered standard errors.
Model (1) (2)
Variable Ln_CH4, Ln_CO2
EUA_price -0.003 -0.004*
(0.211) (0.093)
Electricity_price 0.004 0.006**
(0.147) (0.048)
Cons -0.067 1.890***
(0.266) (0.000)
Observations 136 136
R-squared (within sectors) 0.037 0.078
Notes: P-values in parentheses, the significance levels of the tests are indicated by: * 10%,
** 5%, and *** 1%.
20
Model 1 is shown on the left side in Table 4. The p-values associated with EUA_price and
Electricity_price are 0.211 and 0.147, respectively, indicating that we cannot reject the null
hypothesis based on these results. This means that the coefficients of the EUA_price and
electricity_price are not significant. Therefore no further conclusion can be drawn based on
these results.
Model 2 is shown on the right side of Table 4. It displays significant coefficients for
EUA_price and Electricity_price at a critical alpha level of 0.1 and 0.05, respectively, as
evidenced by their p-values of 0.093 and 0.048. The coefficient of EUA_price demonstrates a
negative relationship, while the coefficient of electricity prices shows a positive relationship.
These results suggest that increasing ETS allowance prices decrease carbon dioxide
emissions. Furthermore, the result indicates that electricity prices have a greater positive
effect on carbon dioxide emissions compared to allowance prices, as evidenced by the larger
coefficient of electricity prices.
The foundation chapter showed that a carbon-based cap and trade system has a negative
effect on carbon emissions, i.e., increased prices on allowances reduce emissions. The effect
on methane emissions from carbon-based solutions is ambiguous among researchers. Ravigné
and Nadaud (2023) showed that when carbon emissions were reduced, methane emissions
increased from the extra usage of methane gas, just as described by the rebound effect
(increased consumption overall) and the backfire effect (increased emission in another area).
Key and Tallard (2011) showed with their research and recursive-dynamic partial equilibrium
model that methane emissions are reduced, irrespective of whether there is an exogenous
price setting for carbon dioxide, such as through the use of carbon taxes, or if a cap and trade
system determine a market price. Key and Tallard (2011) proved that a cap and trade system
based on methane would reductively affect methane emissions. Whether a carbon
dioxide-based cap and trade system also affects methane emissions remains unclear. It is
plausible that implementing policies targeting carbon dioxide reductions may result in profit
increases through methane emission reductions. However, this can not be proven from the
results, which raises the question of why the model does not show an effect.
21
6. Discussion
From the foundation, it was shown that there could be some effect from the EU ETS on
methane emissions. However, model 1 fails to produce a significant result when testing
methane emissions, evident by a p-value of the coefficient EUA_price higher than any
reasonable alpha. This higher p-value could be attributed to various factors, including limited
data points, a small effect from the EU ETS, or a wide data spread (Stock & Watson, 2020, p.
154). These factors can influence the statistical significance of the coefficient estimate and
may explain the higher observed p-value. A small effect size combined with a low sample
size produces insignificant results, as small samples cannot detect minor effects. Lastly, the
data could be widespread because only yearly data is used, leading to higher variance. This
means that the result from any model using yearly data might be less reliable. Thus, it might
be better to use monthly data or wait for more data points until reconducting the model.
Even minor effects would be shown with a good model, meaning our model likely needs
improvement. The model could suffer from endogeneity which would affect the p-values.
Finding additional relevant control variables could fix this potential problem.
Another thing that could affect the result is the selection of a country. Different countries may
receive different effects from the EU ETS. A sufficiently large country could influence the
EU ETS by lobbying for their policies, thereby changing the effect, as Ravigné and Nadaud
(2023) explained. It is also essential to pick a sufficiently small country so that its emissions
do not predict the EUA prices. It is reasonable to expect similar results if another Annex 1
country, similar in size to Sweden, was picked. As its solutions to mitigate emissions are
similar, according to Key and Tallard (2011), and a small country is not able to lobby for its
own policies.
Sweden has no specific regulations or policies that should be more effective on the overall
reduction compared to another Annex 1 (industrialised) country. Any policy regulating sector
emissions using exogenous price settings would only change where the emissions will occur
instead of reducing them, as explained by previous research from Goulder and Schein (2013).
The only case where the model would fail to generalise its result to other populations would
be if the country chosen in the model, in this case, Sweden, would have different exogenous
prices on emissions than other Annex 1 countries. This is likely not the case, as argued by
Key and Tallard (2011).
22
Another possible explanation for why the model cannot produce significant results might be
that the model does not include all relevant control variables. It could be the case that other
variables also impact methane emissions.
If relevant control variables are omitted, the EUA price will be biassed, leading to
insignificant results. Since the model creates a significant result when testing for carbon
dioxide, it could be the case that the models need different control variables. Thus there
might be a need for other control variables to successfully capture the effect of the EU ETS
on methane emissions.
Neither Key and Tallard (2011) nor Ravigné and Naudad (2023) have strong enough
arguments, or enough discussion regarding a cap and trade for carbon dioxide, to state that
there is an effect from the EU ETS on methane. The previous research highlights that it is
theoretically possible for the EU ETS to have an effect in any direction. The discussion of
whether the EU ETS affects methane emissions needs further arguments.
Previous research by Key and Tallard (2011) indicates that carbon taxes effectively reduced
methane emissions in the agriculture sector and that it likely is true for other sectors as well.
They, however, also show that a cap and trade system could have different effects within
different sectors. For future research, it might be better to look at the effect on each sector.
One thing to remember when replicating the method is that the result (the respective p-value
and the strength of effect) could change depending on the definition of sectors. The sectors
capture any eventual policy, or unknown differences between the sectors regarding emissions,
as long as they are constant over time. Controlling for sectors with the fixed effect model
makes the model more accurate by isolating the effect of the EUA price. When replicating the
model, the result could change depending on the definitions of sectors for the country used in
the model. The Swedish Environmental Protection Agency (2022) had separate classes for
national and international transport, for example, and differences in which industry belongs to
what sector may be of discussion. The fixed effect model assumes that the differences are
constant over time. However, policies may change during the time period tested, which
means the result is less reliable. This is a limitation of the model since including all relevant
control variables is impossible (Löfgren et al., 2013).
Due to no inclusion, some of the sectors in the model should not be affected by the EU ETS.
However, an idea of why they could be affected by the EU ETS is that non-included sectors
23
may still trade voluntary carbon credits. The price of voluntary carbon credits likely
correlates with the EUA price, although this needs to be tested. Since methane is price elastic
to the CO2-eq price (Key & Tallard, 2011), the carbon credits may affect methane emissions
even if the sector is not included in the EU ETS. The idea is exciting, but the data for prices
on voluntary carbon credits is difficult to gather.
The electricity price is statistically significant, as shown in Model 2. It can therefore be
deemed a relevant control variable as it can explain the variation in carbon dioxide. The
question then arises about why it is not significant when testing for methane emissions.
Perhaps electricity is not a good choice in predicting methane emissions. The model also
shows that electricity prices have a greater effect on carbon dioxide emissions than the EUA
price. This is interesting to remember since this could mean that the EU ETS is relatively
ineffective at reducing carbon dioxide emissions. The results suggest that electricity prices
are essential to keep low as an increase in electricity prices increases carbon dioxide
emissions. For the thesis, the magnitude of the effect of the electricity price in comparison
with the EUA price might not be exciting. However, in a discussion regarding global
warming, it is. Since higher electricity prices increase emissions, the discussion of electricity
prices may be more important to discuss than the effect of the EU ETS.
40% of the EU electricity production is from fossil fuels, and 20% is from natural gas
(European Council, 2023). Methane is a type of natural gas, which means that a key to
reducing methane emissions may be keeping energy prices low so that companies do not
profit from selling emission-intensive electricity.
As explained previously by Key and Tallard (2011), the effect of the EU ETS could differ
between sectors. This could explain why the model yields an insignificant result. Our model
only examines whether there is an effect from the EU ETS on methane emissions in Sweden
overall and not if there are any sector-specific effects. This means that the possible sectoral
differences are not shown in the model. Adding interaction variables for the sectors could
capture these sector-specific effects and thereby yield a more accurate result.
Another thing to consider when discussing the sectors is the law of demand concept. There
may be situations where the goods cannot be substituted (due to inherent need, such as within
pharmaceuticals). The inability to substitute goods means that the effect of the EU ETS could
be limited for some sectors. This further explains why the EU ETS would affect different
sectors differently.
24
Agriculture, as an example, is a sector in which there are plenty of substitutes, whereas
medicine or transport has fewer options than agriculture. It could also be argued that the
effect of EUA price does not differ between sectors, as long as the emissions mitigation in the
sectors is done by the same method (Key & Tallard, 2011). Key and Tallard highlight how the
effect of a policy depends on what solution is used to decrease the emission. It is plausible
that the sectors have different solutions to decrease their emissions. Therefore, the effect of
the EUA price is likely different between sectors. This gives further basis for using
interaction variables in the model.
The method section mentioned a risk of underestimating or falsely estimating the effect of
EUA price on emissions due to different effects between sectors. This could be why the result
is insignificant and might be mitigated using interaction variables which should improve the
model's accuracy.
The model's accuracy should be sufficient because it produces a significant result when
testing for carbon dioxide, making it challenging to explain why it does not yield a significant
result for methane emissions. Some explanations could be derived from the variables, the
underlying data or the method. Three effects describing effects from one area to another were
mentioned in the foundation chapter. The previously mentioned research papers validated the
effects of these three theories, either in the way that there is a clear negative impact or a
positive. The result in this thesis is insignificant, i.e., we can not determine if there is an
effect. This could be due to the three effects impacting at the same time. There could be a
positive spillover effect, which means that the EU ETS impacts and reduces methane
emissions as carbon dioxide emissions are reduced. At the same time, the carbon dioxide
reduction may backfire and lead to increases in both carbon dioxide emissions and methane
emissions in another area (distinctively different from the first one by locality or sector).
There could also be a rebound effect, that the reduced carbon dioxide or methane emissions
from the EU ETS are rebound and found in other countries. The situation is likely complex,
capturing the effects of the EU ETS is therefore challenging. This gives arguments to further
limit the model to specific sectors, as this would make the situation of different effects less
complex.
Besides the complex situation, the next thing to examine further would be the overall model
and possible limitations. The model is built upon OLS regression, which is also the case for
previous research, such as the research by Löfgren et al. (2013). A better method could exist
25
to model the problem, but to our knowledge, we have yet to find one. Further work could
show this concept using another more complex model, which perhaps could model complex
systems better.
A limitation of the thesis is that the discussion of whether the EU ETS impacts carbon
dioxide is not highlighted. Previous research has discussed whether the EU ETS affects
carbon dioxide, and our model produces significant results, but they are not discussed
thoroughly. Further research on what effects the EU ETS has on carbon dioxide would
contribute to the knowledge on the subject.
Lastly, it is essential to carefully review and be critical when reading research papers. Basing
a model on previous assumptions may be a risk leading to poor or incorrect conclusions. One
of the key things we are critical of with Key and Tallard's (2011) findings is that they assume
that all Annex 1 countries would use the same mitigating solutions (a carbon tax). If this is
false, they would likely receive different effects from the cap and trade system. Therefore,
generalising the model's result to other populations would be difficult. Determining the effect
of carbon-based financial solutions on methane emissions is complicated. The previous
research has presented clear but divided information. Key and Tallard (2011) state that
methane emissions will be reduced in agriculture when the price of carbon dioxide emission
increases. Goulder and Schein (2013) show that increasing carbon dioxide prices are
ineffective at reducing methane emissions in the energy sector. Lastly, Ravigné and Nadaud
(2023) state that reducing carbon dioxide emissions will backfire, leading to increased
methane emissions. This further incentivises testing sector-specific effects, as the previous
research shows that different sectors would receive different effects from a cap and trade
system.
Fewer sectors could be tested with interaction variables to ensure that any sector-specific
effects are noticed. However, the lack of data on methane emissions on a sectoral level makes
future research harder. Regardless, there is a need for more research within the area to
determine the possible effects of the EU ETS.
26
7. Conclusion
The purpose of this thesis was to determine if the price of allowances within the EU ETS has
an impact on methane emissions. The model cannot show that the EU ETS impacts methane
emissions as the model produces an insignificant result. Therefore, the research question
remains unanswered. However, the model produces significant results when testing for
carbon dioxide emissions. The model shows that EUA prices have a negative effect on
carbon dioxide emissions and that electricity prices have a larger positive effect. Conducting
a new model, using interaction variables, more control variables, or a different data set might
be the solution to get a significant answer. Previous research has shown that there likely is an
effect on methane emissions when introducing other financial methods, such as a carbon tax.
However, the direction and magnitude of the effect have been up for discussion.
Some of the research discussed in this thesis indicates a negative effect, such as reductions in
carbon dioxide leading to increased methane emissions. Other research points in the opposite
direction and indicates a positive effect, i.e., that reductions in carbon dioxide emissions also
lead to reductions in methane emissions. Therefore, It is impossible to answer the research
question solely based on previous research. It would be important to test sector-specific
effects because previous research states that financial solutions may have different effects
depending on the sector.
The lack of data on methane emissions is a serious issue that makes researching the area
challenging. Policymakers must acknowledge this lack of information in order to enable
future research.
Policymakers should also be careful when considering the effect of carbon policies. There
could be spillover, backfire and rebound effects, which are challenging to model. Considering
the three effects is essential since adding exogenous prices does not ensure emission
reductions. The best way to reduce one type of emission is to tie the financial solution
specific to that specific emission and that the solution is global. If the EU ETS does not
reduce methane emissions, there could be a need to introduce a new financial solution
specifically targeting methane.
The Paris Agreement states that global warming needs to be limited to a maximum of 2°C by
2050 (United Nations, 2023a). The GWP100 bases the warming effect of gases on a 100-year
27
timeframe. In this timescale, methane is calculated to have a warming effect that is 25 times
greater than carbon dioxide (Eurostat, 2017). However, methane emissions have a global
warming effect of 80 times greater than carbon dioxide during the first 20 years (UNEP,
2022). Therefore it is essential to reduce methane emissions to reach the goals set out by the
Paris Agreement.
28
Reference list
Energimarknadsinspektionen. (n.d.). Elområde. Available at:
https://ei.se/konsument/el/sa-har-fungerar-elmarknaden/elomrade#h-Sverigedeladesinielomra
den [Accessed 9 May 2023].
European Commission. (n.d.a). Development of EU ETS (2005-2020). Available at:
https://climate.ec.europa.eu/eu-action/eu-emissions-trading-system-eu-ets/development-eu-et
s-2005-2020_en. [Accessed 19 March 2023].
European Commission. (n.d.b). EU Emissions Trading System (EU ETS). Available at:
https://climate.ec.europa.eu/eu-action/eu-emissions-trading-system-eu-ets_en. [Accessed 18
March 2023].
European Commission. (n.d.c). Reducing emissions from the shipping sector. Available at:
https://climate.ec.europa.eu/eu-action/transport-emissions/reducing-emissions-shipping-secto
r_en#faq [Accessed 22 March 2023)
European Council. (2023). Infographic - How is EU electricity produced and sold? Available
at:https://www.consilium.europa.eu/en/infographics/how-is-eu-electricity-produced-and-sold/
[Accessed 12 March 2023].
Eurostat. (2017). Glossary: Carbon dioxide equivalent. Available at:
https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Glossary:Carbon_dioxide_e
quivalent. [Accessed 12 March 2023].
29
Formas. (2023). A sustainable transformation for climate action in a changing world.
Available at:
https://formas.se/en/start-page/archive/calls/2023-03-21-a-sustainable-transformation-for-cli
mate-action-in-a-changing-world.html [Accessed 15 May 2023]
Goodwin, N., Harris, J. M., Nelson, J. A., Rajkarnikar, P. J., Roach, B., & Torras, M. (2022).
Macroeconomics in context. Routledge.
Goulder, L.H., & Schein, A.R. (2013). Carbon taxes versus cap and trade: a critical review.
Doi: 10.1142/S2010007813500103
Hayes, A. (2022). What Is the Law of Demand in Economics, and How Does It Work?.
Investopedia. Available at: https://www.investopedia.com/terms/l/lawofdemand.asp.
[Accessed 15 May 2023]
Key, N., & Tallard, G. (2011). Mitigating methane emissions from livestock: a global analysis
of sectoral policies. Climatic Change, 112(2), pp.387–414. Doi: 10.1007/s10584-011-0206-6
Lag (2019:504) om ansvar för god forskningssed och prövning av oredlighet i forskning.
(SFS 2019:1150). Utbildningsdepartementet.
https://www.riksdagen.se/sv/dokument-och-lagar/dokument/svensk-forfattningssamling/lag-2
019504-om-ansvar-for-god-forskningssed_sfs-2019-504/
Löfgren, Å., Wråke, M., Hagberg, T. and Roth, S. (2013). Why the EU ETS needs reforming:
an empirical analysis of the impact on company investments. Climate Policy Doi:
10.1080/14693062.2014.864800
30
NASA. (2011). What Are Climate and Climate Change? Available at:
https://www.nasa.gov/audience/forstudents/5-8/features/nasa-knows/what-is-climate-change-
58.html [Accessed 9 May 2023].
Nordhaus, W.D. (2007). To Tax or Not to Tax: Alternative Approaches to Slowing Global
Warming. Review of Environmental Economics and Policy. Doi: 10.1093/reep/rem008
Nordpool Group. (2023). Day-ahead prices. Available at:
https://www.nordpoolgroup.com/en/Market-data1/Dayahead/Area-Prices/ALL1/Yearly/?view
=table [Accessed 24 April 2023]
Perloff, J.M. (2021). Microeconomics : theory and applications with calculus. Harlow,
England. Pearson.
Ravigné, E., & Naudad, F. (2023). Working Paper, Can a carbon tax increase emissions?
Available at:
https://www.centre-cired.fr/wp-content/uploads/2023/01/cired_wp_2023_89_ravigne_nadaud
.pdf [Accessed 15 May 2023].
Siping, J., Weishi, Z., Ying, X., Yujian, W., Wendai, L., Bashir, A., Shixiong, W. (2021).
Analysis of the spillover effect between CO2 and other pollutants in China.
Doi:10.1016/j.ecolind.2021.107988
Stock, J.H., & Watson, M.W. (2020). Introduction to econometrics. Harlow, Pearson
Education Limited.
31
The United nations’ Sustainable Development Solutions Network. (2022). Spillover effects.
Available at: https://www.unsdsn.org/spillover-effects [Accessed 20 April 2023].
The Swedish Environmental Protection Agency. (2022). Sveriges utsläpp och upptag av
växthusgaser. Available at:
https://www.naturvardsverket.se/data-och-statistik/klimat/sveriges-utslapp-och-upptag-av-vax
thusgaser/ [Accessed 22 April 2023]
Umweltbundesamt. (2019). Rebound effects. Available at:
https://www.umweltbundesamt.de/en/topics/waste-resources/economic-legal-dimensions-of-r
esource-conservation/rebound-effects [Accessed 7 April 2023].
UNEP. (2022). What’s the deal with methane?. Available at:
https://www.unep.org/news-and-stories/video/whats-deal-methane. [Accessed 20 April 2023].
United Nations. (2023a). The Paris Agreement. Available at:
https://www.un.org/en/climatechange/paris-agreement. [Accessed 4 April 2023]
United Nations. (n.d.) Common Metrics. Available at:
https://unfccc.int/process-and-meetings/transparency-and-reporting/methods-for-climate-chan
ge-transparency/common-metrics. [Accessed 4 April 2023]
Vattenfall. (2021). Därför varierar elpriset. Available at:
https://www.vattenfall.se/fokus/tips-rad/darfor-varierar-elpriset/ [Accessed 15 May 2023].
32
World Bank. (2020a). World Bank Open Data. Available at:
https://data.worldbank.org/indicator/EN.ATM.CO2E.KT?locations=SE [Accessed 9 May
2023]
World Bank. (2020b). World Bank Open Data. Available at:
https://data.worldbank.org/indicator/EN.ATM.METH.KT.CE?locations=SE [Accessed 9 May
2023]
World Bank. (n.d.) What is carbon pricing?. Available at:
https://www.worldbank.org/en/programs/pricing-carbon [Accessed 6 April 2023]
World Bank. (2023). World Bank Open Data. Available at:
https://carbonpricingdashboard.worldbank.org/map_data [Accessed 2 May 2023]
33