Corsia Carbon Emitions Aviation

Transportation Research Part D 97 (2021) 102839
Contents lists available at ScienceDirect
Transportation Research Part D

journal homepage: www.elsevier.com/locate/trd
Would CORSIA implementation bring carbon neutral growth in

aviation? A case of US full service carriers☆
Ankit Sharma a, Suresh Kumar Jakhar b, Tsan-Ming Choi c, *
a
Indian Institute of Management Amritsar, Government Polytechnic Campus, Polytechnic Road, Amritsar 143105, Punjab, India
b
Indian Institute of Management Lucknow, IIM Road, Prabandh Nagar, Lucknow 226013, Uttar Pradesh, India
c
Department and Graduate Institute of Business Administration, College of Management, National Taiwan University, Roosevelt Road, Taipei 10617,
Taiwan
A R T I C L E I N F O A B S T R A C T
Keywords: In this study, we develop an econometric model to identify and investigate the determinants of
CO2 Determinants carbon emission by full-service airline carriers for international transportation. The econometric
Full service carriers model explores the CO2 emission determinants for the three largest full-service carriers in the US
Panel data regression
aviation sector. The econometric model uses the fixed effect panel data regression technique to
Simulation
investigate the causal relationship between aviation carbon emission and its determinants. Based
Carbon neutral growth (CNG)
CORSIA on the significant determinants and major technological innovation-based factors, we further
design a simulation model to gauge the impact of each attribute on total carbon emissions of the
airlines when the Carbon Offsetting and Reduction Scheme in International Aviation (CORSIA) is
implemented. The findings suggest that the adoption of alternative biofuels and CORSIA imple
mentation are necessary steps towards a carbon neutral future. Useful insights on route config
uration and adoption of fuel-efficient aircraft fleet are also discussed.
1. Introduction
Climate change and its impacts on the environment around us are a reality today. The aviation industry happens to be one of the
largest carbon (CO2 ) emitters. In fact, it contributes to around 2 per cent of total global (CO2 ) emitted annually (Ansell and Haran,
2020), which is equivalent to approximately 900 million metric tons of (CO2 ) annually. In the decades to come, there is thereby a
pressing need to transit from conventional jet fuels towards hydrogen and other alternative biofuels in order to meet the carbon neutral
growth objectives. Without any policy intervention and biofuel adoption, the emission levels would go beyond control (Ansell and
Haran, 2020). Airlines therefore need to seriously implement carbon emission mitigation schemes and operate in a more responsible
manner.
This study deals with issues that airlines would face when the much needed Carbon Offsetting and Reduction Scheme in Inter
national Aviation (CORSIA) is implemented. To understand the impact of CORSIA implementation, we need a model that could capture
the essence of real-life situations, depicting the issue of carbon emission in the aviation industry. To develop such a model for
☆
We sincerely thank the editors and reviewers for their helpful and constructive comments on our paper, especially during the pandemic time.
We wish them all the safety from the COVID-19 virus. We sincerely thank Bureau of Transportation Statistics, United States Department of
Transportation and MIT’s Global Airline Industry Program for keeping their detailed databases updated at regular intervals in time.
* Corresponding author.
E-mail addresses: ankits@iimamritsar.ac.in (A. Sharma), skj@iiml.ac.in (S.K. Jakhar), tmjchoi@gmail.com (T.-M. Choi).
https://doi.org/10.1016/j.trd.2021.102839
Available online 23 June 2021

1361-9209/© 2021 Elsevier Ltd. All rights reserved.
A. Sharma et al. Transportation Research Part D 97 (2021) 102839
experimentation, at first, we need to understand the factors that play a crucial role in total carbon emission generation. It is only then
can we design and analyze the system to understand the optimal strategies to reduce carbon footprints. Once the policy is effectively
implemented, airlines not only would have to think about their profit margins, but alsoconsider how to minimize their carbon foot
prints while operating in international skies. There are studies that either investigate some crucial determinants of emission or talk
about the impact of emission offsetting policies, seperately but none has done it together. In order to bridge this gap, this study aims at
answering the following questions:
• What are the magnitudes and signs of the critical factors for full-service carriers in the international skies that go on to play a
significant role in the total amount of carbon emissions by an airline per se?
• Should airlines operate more medium haul flights than long haul flights under CORSIA to curtail the total international carbon
emissions?
• Which aircraft combinations, technologies and changes would help airlines achieve a carbon neutral growth?
To answer these questions, this study has chosen to adopt two methodologies. The answer to the first question is determined using a
systematic literature review, followed by collation of secondary data for the three leading full-service carriers of the U.S. After that, the
data are analyzed using the panel data regression technique, through which significant factors are identified. The answers to the next
two questions are determined using a simulation model, which in turn is based on the critical factors that have been identified earlier,
along with some additional technological development factors that have been identified in the process of developing the conceptual
model. The simulation model analyzes the total carbon emission patterns for the next 15 years, starting from 2021 till 2035, when
CORSIA is implemented. The systems dynamics simulation modelling technique is used for the analyses via the Vensim simulator.
1.1. The US airline industry overview
The US air transportation market is the largest in the world. Herein, both hub and spoke network and point to point network co-
exist, whereby it has both full-service carriers and low-cost feeder carriers that fulfill the passenger demands across the origin and
destination pairs (Zou et al., 2012). The feeder carriers generally operate on smaller aircraft between non-hub OD pairs, focusing on
Fig. 1. Performance attributes of three largest full service carriers in US market (A).
2
high frequency services on short haul routes (Reynolds-Feighan, 2018). Notably, all the analyses presented in this paper would be
based on the three largest full-service airlines in the U.S., that include the American Airlines (AA), Delta Airlines (DA) and United
Airlines (UA). Using the official database of the Bureau of Transportation Statistics, monthly data points for the three airlines were
collected for 17 years, starting from 2003 till the end of 2019. The growth of the three airlines over this period was assessed from
various performance attributes shown both in Fig. 1 and Fig. 2. The general trend of all the attributes shows a year-on-year cyclic trend
in growth for each player. As Belobaba et al. (2015) discuss, the profits in the global airline industry have been cyclical and extremely
variable due to many global events that directly impact the airlines. The industry posed consecutive loses during 1990–1993 due to the
Gulf War and recession. It recovered in late 1990s, and plunged into dramatic economic losses between 2001 and 2005 due to 9/11
Attack in the USA in 2001. The global recession from 2008 onward hit the industry once again, with highly volatile financial markets.
All these events make the indicators shown in Fig. 1 and Fig. 2 look cyclical. Abnormal upsurges in the Available Seat Miles (ASM), the
Revenue Passenger Miles (RPM) and the international flights shown in Fig. 1 (a, b and c) for the United Airlines from 2012 are due to its
merger with Continental Airlines.
The load factor for the industry as a whole, and for each individual airline has steadily improved annually, thereby signifying an
improved productivity and high occupancy per flight, which suggests better resource utilization is achieved. This fact may be un
derstood from the components (b) and (c) of Fig. 1, where there is no sudden upsurge in total international flights, while substantial
improvements are seen in RPM. This in turn signifies improvement in occupancy per flight, as validated by the load factor data. All
these factors go on to affect the carbon emission levels each year, with which we would establish the suitable hypotheses. Although
there are several studies that have analyzed the operations and economic performance of the US airline industry, this study is very
different from them, because it focuses exclusively on the environmental aspect. Fig. 2 suggests a sudden spike in Delta Airline’s
indicators from 2010 onward. This is due to the completion of the acquisition of Northwest Airlines into the operations of Delta
Airlines.
1.2. Contribution statement and paper structure
An important novelty of this work comes from the uniqueness in the way CORSIA is explored. Moreover, it is the first study that
Fig. 2. Performance attributes of three largest full service carriers in US market (B).
3
employs panel data to determine the crucial influencers of carbon emissions for full-service carriers in international air travel.
Importantly, it utilizes these determinants in simulating the impact of CORSIA implementation under pessimistic, optimistic and
realistic scenarios. Furthermore, this study also discusses the importance of adopting new aircraft fleets and alternative green fuels in
order to achieve carbon neutrality when CORSIA is implemented. The importance of higher proportion of medium haul flights in the
CORSIA era would serve as a key strategic consideration for airlines.
The structure of the paper is as follows: Section 2 examines the extant literature on emissions in the airline industry, and the factors
that affect the emissions for an airline. There is also a discussion on the salient features of CORSIA, and about environmental emission
control. Section 3 is dedicated to the panel data regression modelling. Section 4 deliberates on the development of causal loop diagram
(CLD) and the stock flow diagram (SFD). Section 5 discusses the model setup and the simulation scenarios to be analyzed. Section 6
discusses the results obtained by the simulation runs for different scenario sets. Section 7 highlights the managerial implications for the
airlines and a discussion on the strategic importance of findings for the decision makers. The final section discusses the scope for future
work.
2. Literature review
The aviation sector has gained substantial popularity in the past few decades due to the availability of flight services among almost
every origin and destination (OD) pair, either directly or via connecting flights. This has resulted in an increase in commercial flights,
and consequently in increased global greenhouse gas (GHG) emissions. Growing concerns for climate change has led to stringent global
norms that have been introduced to curb GHG emissions, which in fact has burdened the airlines with additional taxes and penalties
(Zou et al., 2012). Notably, the global emission curbing schemes in the international skies follow fuel consumption-based methods to
calculate the emission levels. This has almost made it imperative for airlines to curb their fuel consumption by employing more fuel-
efficient fleets, thereby retiring the older fuel-inefficient ones. Airbus, the largest aircraft manufacturer, has announced their plans of
bringing zero emission commercial aircraft in operations by 2035 (Ambrose, 2020). However, till the time when zero emission aircraft
becomes a reality, we must recognize that with sevenfold increase in air traffic, the annual emissions, as predicted by the International
Civil Aviation Organization (ICAO), would increase to 400–600% by 2050, as compared to their levels in 2010, if no emission miti
gation policy is implemented by then (Cui and Li, 2017).
2.1. Aviation and the carbon emission
Several studies are carried out every year to explore the impact that the airline industry has on the environment. As the pressure
towards environmental well-being is on the rise, and the market is becoming more and more volatile (with increasing volatility in the
fuel price), airlines have to manage their profit margins, while taking care of their carbon footprints (Miyoshi and Fukui, 2018).
Concerns over carbon emissions and volatile fuel prices have potential impacts on airlines operations, as fuel cost directly corresponds
to around 15–25% of the total operating costs (Brueckner and Zhang, 2010). Cui and Li (2017) evaluate the fuel efficiency of 29 global
carriers under the Carbon Neutral Growth (CNG) 2020 and the authors suggest that the pollution causing inputs need to be more
efficient if we want to mitigate the overall emissions. Some of the key studies that examine fuel efficiency and carbon emission, cover
aspects like the pollution abatement cost (Cui et al., 2018), biofuel’s impact on economic and emissions goals (Winchester et al., 2013),
optimal pricing strategy when carbon taxation is in place (Cui, 2019), and the impact of technological innovations on airline emissions
(Grampella et al., 2017). Chao (2014) assesses the carbon emission cost for air cargo services in US. Rotaris et al. (2020) study the
consumer’s willingness to pay for the carbon offsetting requirements in Italy. The work suggests that passengers are willing to pay for a
carbon neutral growth 2020 goal in the range of 12–38 lb per ton and 14–66 lb per flight. Yuen and Zhang (2011) study the impact of
unilateral emission regulations on domestic airlines and discuss the concern of competitive advantage to foreign airlines only. Evans
(2014) highlights the positive impact of network change on airline aviation. Arul (2014) suggests methods to monetize the variation in
load factor and emissions for US domestic carriers. Lu and Shon (2012) evaluate the passengers’ willingness to pay for the carbon
offsets. Liu et al. (2017a) propose a global Malmquist carbon emission performance index to measure the dynamic changes over time.
They suggest that technological progress would substantially reduce emissions in the airline industry. Some other studies investigate
the impact of carbon taxation fee on passenger demand. Qiu et al. (2020) discuss the carbon tax incentive policy for aviation emission
reduction. Further, Pagoni and Psaraki (2016) suggest that air passenger demand can fall between 2% to 21% depending on the carbon
credit rate per ton. Liu et al. (2020) determine the drivers of carbon emissions using a two staged decomposition method based on
efficiency. Chao et al. (2019) study the impact of CORSIA implementation on the US airline industry. They suggest that there is a 3.5%
chance that the US aviation sector would reduce GHG emissions by 37.5–50% by 2050.
IATA airlines and various stakeholders in the aviation sector agree on three international aviation’s climate goals, viz. freezing the
global emission levels from 2020 (CNG 2020), improvement in fuel efficiency of around 1.5% during 2009 to 2020 and global emission
reduction to 50 percent of 2005 levels by 2050 (Maertens et al., 2020). CORSIA’s road-map to achieve these goals focuses on four
measures, viz. adoption of biofuels and new technology aircraft, modern air traffic management, fuel saving flight operations and a
global market based measure (CORSIA). Carlsson and Hammar (2002) suggest that incentive based emission trading schemes are
better regulations than the conventional comman and control schemes such as limiting cap and technological standards. However they
are complex in terms of implementation. In-depth comparisons of the incentive based emission schemes and the command and control
based schemes are covered in Carlsson and Hammar (2002), Toke (2008), Carlsson and Hammar (2002), Vespermann and Wittmer
(2011). The two biggest emission control schemes, namely the European Emission Trading (ETS) scheme and CORSIA, are both market
based approaches unlike the conventional command and control schemes rolled out in the past. Formerly, due to the absence of a global
4
A. Sharma et al.
Table 1
Literature review comparing the works addressing the issue of carbon emissions determinants.
Reference Region Period of Context Factors Investigated Regression
Study Technique
FUELt FUEL/Gt STAGEt DEPt SIZEt RPMt ADAYSt LFt Biofuels Aircraft Emission
Fleet Tax
Givoni and Rietveld Global - international - - - ✓ ✓ - - ✓ - ✓ - -

(2010)
Hofer et al. (2010) United 2004 domestic - - - - - - - - - - ✓ -

States
Scheelhaase et al. Global 2006–2012 international ✓ - ✓ - ✓ ✓ - - - - - -
(2010)
Yamaguchi (2010) Japan, US 1998 domestic - - ✓ ✓ ✓ - ✓ - - - Difference in
Difference
Dray (2013) Global - international - - - - - - ✓ - - ✓ - -
Winchester et al. United 2020 domestic - - - - - - - -
5
✓ ✓ ✓ ✓
(2013) States
Brueckner and United 1995–2015 domestic and - ✓ ✓ - ✓ ✓ ✓ ✓ - - - Ordinary Least
Abreu (2017) States Square
international
Fukui and Miyoshi United 1995–2013 domestic and ✓ ✓ - - - - - - - - ✓ Quantile
(2017) States Regression
international
Grampella et al. Global - international - - - - - - - - - ✓ - Ordinary Least
(2017) Square
Liu et al. (2017a) China 2007–2013 domestic - - - - - - - - - ✓ - Panel Data
Lo et al. (2018) Italy 1997–2011 domestic - ✓ ✓ - ✓ - - - - - - Panel Data
Transportation Research Part D 97 (2021) 102839

Sibdari et al. (2018) United 2005–2015 domestic ✓ - - ✓ ✓ - - ✓ - - - Two Stage Least
States Square
Chao et al. (2019) United 2005–2050 domestic - - - - - - - - ✓ ✓ - -
States
This Paper United 2003–2035 international ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ - Fixed Effect
States Panel Data
Table 2
Literature review comparing the works addressing the issue of carbon emissions in aviation industry.
Reference Region Period of Policy Studied Predicitive Dataset Accomplishments
Study modeling
EU CORSIA
ETS
Givoni and Global - - - - ICAO Engine Exhaust Adjusting service frequency and
Rietveld Database increasing the aircraft size to offer
(2010) equivalent seating
capacity will aggravate local pollution but
reduce the climate change impact
Hofer et al. United 2004 - - - BTS, DB1B Studies the impact of emission taxes on
(2010) States airline emissions.
Scheelhaase et al. Global 2006–2012 ✓ - Mission OAG, BADA Impact of EU-ETS on competition among
(2010) Analysis the EU and non EU airlines
Yamaguchi Japan, 1988–2007 - - - BTS, AATS Japan Outlook of Japanese airlines towards
(2010) US voluntary emission reduction schemes are
studied.
Dray (2013) Global - - - OAG Global Fleet Variations in retrofits and retirement age
Data have had little effect on total global
emissions.
Winchester et al. United 2020 - - Simulation Global Trade solves economy-wide model for energy
(2013) States Analysis systems and economic activity and an in-
depth
Project database, partial equilibrium one for the aviation
industry to examine the emission and
economic
International Energy impact goals of biofuels by the US Federal
Agency Aviation Administration (FAA).
Brueckner and United 1995–2015 - - - Bureau of Paper describes regression results linking
Abreu (2017) States Transportation an airline’s total fuel consumption to
crucial
Statistics, (BTS) variables and shows how emissions and
fuel consumption depend on these
variables
Fukui and United 1995–2013 - - - Environmental Impact of carbon taxation on airline
Miyoshi States Protection Agency carbon emission and fuel consumption is
(2017) studied.
Grampella et al. Global - - - - Developed by authors Effects of incremental as well as
(2017) substantial innovations on noise levels
and aviation
emissions among aircraft/engine blend of
Airbus A320 and Boeing B737 families.
Liu et al. (2017a) China 2007–2013 - - new Statistical Data on Changes in Revenue Ton Kilometers
generalized Civil (RTK) is the biggest factor responsible for
PDA increased
method Aviation of China civil aviation CO2 emissions and changes
in potential energy intensity is crucial in
decreasing CO2 emissions for most
airlines.
Lo et al. (2018) Italy 1997–2011 - - - OAG Dataset, Aircraft size increases total emissions, but
it brings emissions per ASK down,
airport specific and total emissions increases with the
route distance, it decreases the emission
per ASK, implying that CO2 is less of a
problem for long-haul connections.
Sibdari et al. United 2005–2015 - - - RITA, BTS Airlines adjust both aircraft sizes and
(2018) States flight frequency to maintain load factors
and
manage airline capacity to account for
fluctuations in fuel cost and passenger
demand.
Chao et al. (2019) United 2005–2050 - ✓ Simulation Bureau of The study utilizes a combined model of
States Transportation multi-feedstock sustainable aviation fuels
(SAFs)
Statistics, (BTS) and airlines operations to signify
decisions of several stakeholders, like bio-
refineries,
farmers, policymakers, and airlines.
This Paper 2003–2035 - Simulation
(continued on next page)
6
Table 2 (continued )
Reference Region Period of Policy Studied Predicitive Dataset Accomplishments
Study modeling
EU CORSIA
ETS
United ✓ Bureau of This work idenfies the critical

States Transportation determinants of carbon emission for the
full service carriers
Statistics, (BTS) in the US and simulate the impact of
CORSIA implementation on the carbon
footprints of
the airlines.
policy on aviation emissions, the EU included international aviation emissions under the ongoing ETS which did not gain consensus
among various airlines, countries and trade associations (Preston et al., 2012). Some of the key studies to explore ETS and its impact on
the aviation sector are Albers et al. (2009); Anger (2010), Scheelhaase et al. (2010), Vespermann and Wittmer (2011) and Nava et al.
(2018). A global need to shift policy governance from the EU to ICAO arose in the later years in the hope of achieving greater emission
reduction. This gives birth to CORSIA.
2.2. Factors affecting the GHG emissions in the airline industry
Several factors are directly or indirectly related to the level of GHG emission from an airline. A few studies uncover various factors,
but only a limited amount of literature empirically establishes the magnitude of the effect that a factor per se has on the emissions. The
overall emission calculations without having an empirically established relationship with the key determinants may not be of much
help in designing a suitable Emission Trading Scheme (ETS). For example, consider the study of Lo et al. (2018), who calculate the
carbon emissions based on the average distance traveled. They conclude that GHG emission is less significant for long-haul connections
as the emission per available seat kilometers (ASK) decreases with increasing route distances. Further, they find that technical progress
decreases CO2 emissions per ASK with an estimated elasticity equal to - 0.06%. Importantly, some of the factors that have been
supported by prior studies and that are known to affect carbon emissions, have been shortlisted in our study for developing the
econometric model. The aim of the econometric model would be to find suitable magnitude and sign for those determinants.
Facing higher demand, airlines typically tend to respond by increasing their flight frequency rather than opting for a bigger aircraft
(Givoni and Rietveld, 2009). In order to make their operations profitable, airlines focus on increasing the load factor. However, if the
market keeps growing and the current competition levels are maintained, airlines must increase flight services, rather than deploying
bigger aircrafts (Givoni and Rietveld, 2010). Sibdari et al. (2018) analyze the impact that the jet fuel cost renders on airline frequency,
the aircraft size and the load factor. With a rise in passenger demand, their findings recommend smaller aircraft with higher frequency
and higher load factor. On the other hand, as the fuel price increases, the airlines would possibly prefer larger aircraft in order to reduce
the per passenger emission levels. Further in these lines, Morrell (2009) studies the impact of larger aircraft on reduction of the total
carbon emission in the European market. The findings highlight a positive relation among the aircraft size and emission reduction,
especially for long haul flights. Merkert and Hensher (2011) concur with the findings, establishing a positive relationship between
aircraft size and fuel efficiency. In fact, they find that a smaller variation in family of aircraft further improves the fuel efficiency. It is
commonly believed that larger aircraft are cleaner than smaller ones for the environment (Brugnoli et al., 2015).
Further, Liu et al. (2017b) suggest a Revenue Ton Kilometers (RTK) model and reveal RTK to be the largest driver of carbon
emissions in the Chinese aviation sector. Revenue Passenger Miles (RPM) signifies an upsurge in demand for air travel, and an increase
in actual travels signifies more emission, a construct worth testing. Zou et al. (2012) employ ratio-based, stochastic and deterministic
frontier approaches to evaluate fuel efficiency among 15 large carriers (mainline airlines) in the U.S. The authors confirm a strong
correlation among RPM, flight frequency and fuel consumption. Miyoshi and Mason (2009) suggest that the key areas for an airline to
curb its total emissions include employing fuel efficient fleet and improving load factor for reducing per passenger emission. Biofuels
have been a very small part of total fuel consumption (2% of total fuel consumption in 2018); however, it can play an important role in
reducing emissions generated from the conventional jet fuel (Winchester et al., 2013). The study further suggest that under a high
carbon credit pricing scheme with optimistic assumptions regarding the growth of biofuel technologies, 100% fuel consumption in
global aviation would be sourced from biofuels by theearly 2040s. With slow development of biofuel technologies and no carbon price,
biofuels would account for 3% of total aviation fuel use in 2030 and around 37% in 2050 (Winchester et al., 2013). Therefore, in order
to understand the significance of biofuels, we investigate the impacts of their adoption in our simulation model. Table 1 and Table 2
compare studies that address the issue of carbon emissions in the aviation industry. The comparison of key studies helps to highlight
the literature positioning of our work.
2.3. Carbon Offsetting and Reduction Scheme in International Aviation (CORSIA)
The Kyoto Protocol (1997) extends the United Nations Framework Convention on Climate Change (UN- FCCC) to the world level. It
aims at reducing GHG emissions, but there are no clear guidelines for the aviation sector. Since then, the International Civil Aviation
Organization (ICAO) is working on a proper policy framework for regulating emissions in the aviation industry. The 37th ICAO
7
Assembly has agreed on Carbon Neutral Growth (CNG) from the year 2020. This implies that airline emissions post 2020 shall remain
below the emission levels in 2020, which has been kept as the base year for the agreement. Finally, the 39th ICAO Assembly has
formalized the scheme as CORSIA. CORSIA is a carbon offset scheme that uses carbon credits for compliance. Scheelhaase et al. (2018)
summarize the key features of CORSIA. Some of the key features worth noting include the fact that CORSIA is a carbon offsetting
scheme. In other words, the emissions from airline operations have to be offset, either by purchasing carbon credits or by investing the
equivalent amount in some green projects that aim at reducing carbon footprints. Second, offset requirements have been discussed in
the form of routes that are flown; i.e. only those routes that lie between the participating states (CORSIA states, by 2021, 88 members
have agreed to join (ICAO, 2017) would represent the CORSIA routes. Third, all airline operators have to surrender carbon offsets,
regardless of their country of origin, apart from small fleets that emit less than 10,000 tons per annum, or are operating under hu
manitarian/medical operations or are new entrants (less than 3 years old in the market). CORSIA does seem to be very timely, as people
today are emission conscious (Choi et al., 2018), especially when they travel in the domestic market. Interestingly, they seem to be less
concerned about the international skies.
Thus, all the international routes between the CORSIA states would be liable for having a carbon offsetting tax, unless they operate
under humanitarian operations. Thus, in our intercontinental analysis, each intercontinental hub edge would be a CORSIA route. Thus,
each player that emits GHG would eventually need to surrender carbon offsets from the profits they earn or invest the equivalent
amount in some green projects, in or outside the aviation sector.
3. Causal relationship between aviation carbon emission and its determinants
This section provides details of our empirical study, whereby we look to identify the determinants of carbon emission for an airline.
To determine the causal relationship between the total carbon emission from international flight operations (long and medium haul)
and the determinants of emission, we propose certain hypotheses. The aim of this exercise is to gauge the magnitude and sign of these
determinants, so that they may be utilized in developing the simulation model for deeper investigation. Thus, we initially design an
econometric model to determine the impact of certain factors on total carbon emission in international flight operations. Then, we
apply the model to a dataset spanning 17 years, (i.e. 2003–2019), encompassing the three largest full-service carriers in the US.
Notably, our rationale for considering two decades of data (i.e. 2003 to 2019) is the continuous global increase in international air
travel. In fact, these years capture the growth of airlines and the larger aviation sector holistically. Moreover, it also allows us to have
204 data points for each variable per airline, and over 600 data points for the each variable for the sector as a whole within a combined
panel data. Importantly, the last year is restricted to 2019 instead of 2020 due to the unavailability of data points caused by an un
precedented fall in international air travel owing to the ongoing COVID-19 pandemic. With the determinants being identified, and
using 2019 as the base for the latest aviation emission data, we would run a simulation for years 2021–2035 in order to assess the state
of emission upon CORSIA implementation.
3.1. Hypotheses
In this section we present the determinants extracted from extant literature and hypothesize their relationship with total carbon
emissions. Using secondary data obtained for full-service carriers, we establish herein the magnitude and sign of the hypothesized
relationship. The need for proposing the hypotheses is that none of the prior studies shown in Table 1 and Table 2 have determined the
uncovered factors together for full-service carriers in international air travel.
Hypothesis 1. (RH1) The total expense on the fuel purchase i.e. the fuel cost is a positive determinant of the total carbon emissions.
Total fuel expense (TFE) is an indicator of how much fuel is being consumed by the airline in a given time period. It is hypothesized
that the higher the TFE, the more would be the fuel consumption at a given unit fuel cost. The higher the fuel cost per gallon, the more
prudent an airline would be in consuming fuel. This leads us to our second research hypothesis about fuel cost per gallon. Moreover, a
higher fuel cost per gallon may partially translate into a higher ticket fare which ultimately results in lowering the demand. This in fact,
leads us to our second hypothesis on fuel cost per gallon.
Hypothesis 2. (RH2) The fuel cost per gallon is a negative determinant of the total carbon emissions in the international carrier
services.
Similarly, the higher the average stage length, the more the distance is covered by the airline and hence the resulting carbon
emissions would be higher. This brings us to our next hypothesis about average stage length.
Hypothesis 3. (RH3) The average stage length is a positive determinant of the total carbon emission.
Moreover, the distance flown by an airline is the product of average stage length and the total number of departures. Thus, for a
fixed total international distance flown, increasing the number of departures would mean reduction in average stage length, signifying
thereby lesser time in international skies per flight, reducing carbon emissions in the process. This leads us to our next hypothesis.
Hypothesis 4. (RH4) The total number of departures is a negative determinant of the total carbon emissions in the international
skies.
Bigger aircraft in the long haul flights have higher fuel burn rate per km and thus would result in higher emissions. To investigate
this phenomenon, we define our next research hypothesis on aircraft size.
8
Hypothesis 5. (RH5) The aircraft size is a positive determinant of the total carbon emissions.
Similarly, the larger the number of passengers travelling via airlines, the more would be the revenue passenger miles and more
would be the resulting emissions in the environment.
Hypothesis 6. (RH6) The revenue passenger miles (RPM) is a positive determinant of the total carbon emissions in the
environment.
With technological innovations, newer fleets are fuel efficient and reduce carbon footprints to a great extent as opposed to older
fleets. This brings us to our next research hypothesis on aircraft days, i.e. the number of days an aircraft is in service for an airline.
Hypothesis 7. (RH7) The aircraft days (ADAYS) is a positive determinant of the total carbon emissions.
Now, the load factor is a sign of how productively an aircraft is used by an airline. The higher the load factor, the lesser the emission
per passenger. But the total emission should be higher with a higher load factor unlike emission per passenger. This leads us to evaluate
the second last research hypothesis for the load factor.
Hypothesis 8. (RH8) The load factor is a positive determinant of the total carbon emissions in the environment.
Finally, when airlines face higher offsetting charges, they are bound to transfer some portion to the passengers. Thus, the higher
airfares would mean there are either higher offsets or some deterrent towards frequent air travel, possibly reducing the total carbon
emissions. This leads us to investigate our last research hypothesis for average airfare.
Hypothesis 9. (RH9) The average international airfare is a negative determinant of the total carbon emissions in the environment.
Based on these hypothesized factors, we develop an econometric model to check which of these factors significantly affect the total
carbon emission. Notably, the significant determinants would then be used to build our simulation model. In order to investigate these
determinants, we first calculate the total carbon emissions based on the data collected for the total monthly fuel consumption by the
airlines from 2003 to 2019. The fuel consumption calculations are done based on the emission calculations guidelines for international
aviation by ICAO (2017).
3.2. The econometric model
To estimate the magnitude, sign, and significance of the determinants, we develop an econometric model (Odeck, 2019; Miranda
and Oliveira, 2018; Miranda and Oliveira, 2018; Brueckner and Zhang, 2010) based on the panel data we have collected. The cross-
sectional variables represent the three airlines, with each having a time series monthly data for various determinants. Further, we have
noted about 612 observations per variable in the panel data to investigate the significance of the variables in the model. Notably, these
612 data points for each variable have been extracted with no missing observations. Thus, we have a strongly balanced panel data for
investigating the proposed model, which is as follows:
TotalCOit2 = α + β1 FUELit + β2 FUEL/Gt + β3 STAGEit + β4 DEPit + β5 SIZEit + β6 RPM it

(1)
+β7 ADAYSit + β8 LFit + β9 EUETS + β10 AVGFAREit + β11 TIMEt + εit
where, i is the index regarding the three airlines and t is the month. We develop a panel data model with one month as the time duration
between the observations. Table 3 describes the meaning of the symbols in Eq. (1). The model is used to estimate for the unobserved
heterogeneity. The unobserved heterogeneity is captured using the dummy variable TIME given by each month over the complete
17 years horizon. The total CO2 emission is estimated in tonnes for each airline at a given time period. When EU ETS was implemented
Table 3
Description of the variable in the econometric model.
Variable Description
Total COit2 Total carbon emission (Tonnes) by the airline i in the month t
FUELit Total fuel cost ($) incurred by an airline i in the month t
FUEL/Gt Fuel cost per gallon ($ per gallon) in the month t
STAGEit Average stage length (KM) of an airline i in the month t
DEPit Total departures of an airline i in the month t
SIZEit Size of the aircraft of an airline i in terms of number of seats in the month t
RPMit Revenue passenger miles generated by an airline i in the month t
ADAYSit Aircraft days i.e. the number of days an aircraft is in service with an airline i in the month t
LFit Load factor of an airline i in the month t
EUETS Dummy variable for EU ETS policy implementation
TIMEt Dummy variable for time in month
AVGFAREit Average international airfare ($) for the airline i in the month t
9
in 2012 to account for aviation carbon emissions, its impact started to happen on the airline emissions. We incorporate a dummy
variable for this policy implementation which takes a value = 1 from 2012 onward and is zero in prior years.
The dependent variables and their regressors can potentially vary over individuals and time. To gauge these variations, we analyze
the descriptive statistics for these variables. The three types of variances are calculated based on the following expressions:
1 ∑∑
s2O = (xit − x)2 , (2)
NT − 1 i t
1 ∑
s2B = (xit − x)2 , (3)
NT − 1 i
1 ∑∑ 1 ∑∑
s2W = (xit − xi )2 = (xit − xi + x)2 , (4)
NT − 1 i t NT − 1 i t
s2O ≈ s2B + s2W , (5)
where, Eq. (2) represents the overall variance i.e. variation over individuals and time, Eq. (3) represents the market variance i.e.
Table 4
Descriptive statistics of the variables.
Variable Mean Std. Dev. Min Max Observations
Total COit2 overall 1361995 434279.9 442447.6 2305604 N = 612

between 130555.6 1257598 1508378 n=3
within 420971.8 484434.3 2259314 T = 204
FUELit overall 2.19E+08 1.07E+08 2.43E+07 4.93E+08 N = 612

between 2.54E+07 1.90E+08 2.37E+08 n=3
within 1.05E+08 1.36E+07 4.79E+08 T = 204
FUEL/Gt overall 2.136454 0.7193239 0.71 4.43 N = 612

between 0.0897424 2.069951 2.238529 n=3
within 0.7155759 0.6079248 4.327925 T = 204
STAGEit overall 2369.606 364.7129 1285.568 2986.175 N = 612

between 173.7136 2258.568 2569.794 n=3
within 335.9538 1396.606 2815.151 T = 204
DEPit overall 447702.3 143109.5 292139 861482 N = 612

between 21263.03 434463.5 472228.9 n=3
within 142050.7 295953.4 836955.4 T = 204
SIZEit overall 208.7892 7.978881 189 220 N = 612

between 6.496163 202 214.9461 n=3
within 5.956766 195.7892 223.7892 T = 204
RPMit overall 3.40E+09 2.04E+09 5.68E+08 9.91E+09 N = 612

between 1.80E+09 2.34E+09 5.47E+09 n=3
within 1.41E+09 1.60E+07 7.83E+09 T = 204
ADAYSit overall 91707.27 37445.41 42331.26 187632.1 N = 612

between 11121.94 80215.23 102417.9 n=3
within 36325.73 53823.3 176921.4 T = 204
LFit overall 0.8026709 0.0529723 0.6234 0.9304 N = 612

between 0.0142165 0.7863716 0.8125108 n=3
within 0.0516827 0.6169405 0.9466994 T = 204
EUETS overall 0.4705882 0.4995425 0 1 N = 612

between 6.80E-17 0.4705882 0.4705882 n=3
within 0.4995425 0 1 T = 204
AVGFAREit overall 721.1663 229.1872 329.1468 1530.449 N = 612

between 212.1549 524.3213 945.8771 n=3
within 149.9039 383.7725 1305.738 T = 204
10
variation across individual (invariant of time) and Eq. (4) represents the internal variance i.e. variation over time for a given individual.
Further, xit is the variable, xi is the individual mean and x is the overall mean. Eq. (5) represents the decomposition of the overall
variance in terms of market and internal variance. The descriptive statistics of the variables used in the model is shown in Table 4.
3.3. Fixed effect model for panel data regression
We use the fixed effect (FE) model to analyze the relationship between the outcome variables and the predictor within an individual
(here airline). We make two major assumptions while using the FE model as described by Torres-Reyna (2007). First, the assumption of
having a correlation between predictor variables and the entity’s error terms. Importantly, FE removes the time-invariant charac
teristic effects, and allows us to assess the net impact of the predictor variables over the outcome variable. Second, the necessity of the
uniqueness of the time-invariant characteristics within the individual (i.e. airline), and the absence of any correlation with other
individual characteristics. Furthermore, in order to test the correlation between the unique error terms with the regressors, and the
suitability of the FE model over random effect (RE), we use the Hausman test (Hausman, 1978) to confirm whether the FE is the right
model to perform the analysis. The results of Hausman test are tabulated in Table 5.
3.4. Results of the econometric model
The results of the econometric model are tabulated in Table 6. The model is run under the TIME dummy to account for the time
fixed effect. The determinants described in the hypotheses are investigated for their magnitude and sign. As shown in Table 6, we can
see that 6 out of 10 determinants are significant except ADAYSit and LFit at every level. Note that, Dray (2013) discusses a similar
finding based on historical data from 1960 to 2005 regarding aircraft age and its impact on emissions. Dray (2013) also suggests that
aircraft age does not contribute to any significant reduction in total carbon emissions. The statistical result confirms that the model is
fine and all the determinants have coefficients different from zero. Further, the correlation coefficient − 0.5822 suggests that the error
terms ui are correlated with the regressors in the FE model.
Looking at the results and revisiting the hypotheses, we confirm that the total fuel cost FUELit is a positive determinant of emission,
while the fuel cost per gallon FUEL/Gt shows a negative relationship, thereby our hypothesis is supported, and enabling us in the
process to use these determinants with a proper sign. Notably, Lo et al. (2018) expect that the fuel cost per gallon is a negative
determinant, but find it to be positive, albeit unexpectedly. The authors explain the reason behind this unexpected finding as, airlines
transferring the increased fuel cost to the passengers and not pushing towards fuel efficiency. Further, the average stage length
(STAGEit ), aircraft size (SIZEit ), and revenue passenger mile (RPMit ) are positive determinants of emission. The results for (STAGEit )
and (SIZEit ) are similar to Lo et al. (2018)’s findings. The total departures (DEPit ) is also a negative determinant of emission but only at
10 percent significance level. The aircraft days ADAYSt or the age of the aircraft is a positive determinant, but not a significant
influencer of the total emissions. The potential reason being intermittent re-engineering, engine and air frame modifications during the
life cycle of the aircraft (Dray, 2013). Thus, we can avoid using the aircraft fleet age variable in the simulation model.
As expected, the sign of the EU ETS comes out to be negative which signifies that the introduction of emission trading scheme would
reduce the total carbon emissions. Although, as ETS covers only 8 years under the analyzed data, ETS does not turn out to be a sig
nificant determinant. Another important determinant, AVGFAREit comes out to be a negative determinant of the total carbon emis
sions. Brueckner and Zhang (2010) support our results on airfare, as their study suggests that higher emission levels result in higher
emission charges that would drive the airfare upwards and consequently reduce the flight frequency. Further, load factor (LFit ) is a
coming out to be a negative determinant of emission but does not play a significant influencer role in total emission. However, LFit is a
significant influencer when we look at the emission per RPM or emission per ASM. This is consistent with Brueckner and Abreu (2017),
Table 5
Hausman test results.
(b) (B) (b-B) sqrt(diagonal(Vb − VB )) Result
fe re Difference S.E.
FUELit 0.0032729 0.0041983 − 0.0009254 .

FUEL/Gt − 324564.3 − 405419.7 80855.46 693.3894
STAGEit 376.6065 523.3213 − 1.47E+02 .
DEPit − 0.2374663 1.014272 − 1.251738 .
SIZEit 10146.85 − 5219.472 15366.32 .
RPMit 0.0000857 2.39E-06 0.0000833 .
ADAYSit − 0.355993 − 2.919112 2.563119 3.21E-06
LFit − 69408.8 407788.2 − 477197.1 .
EUETS − 22421.66 − 49054.75 26633.09 .
AVGFAREit − 123.6948 − 49.9871 − 73.7077 .
χ2 468.90
p > χ2 0.0000
11
Table 6
Fixed-effect regression model results.
Fixed-effects (within) regression F(10,599) = 2224.32
corr(ui , Xb) = − 0.5822 Prob > F = 0.0000
Total COit2 Estimated Coefficient Std. Err. t p-value [95% Confidence Interval]
FUELit 0.0033*** 0.0001 34.4200 0.0000 0.0031 0.0035

FUEL/Gt − 324564.3000*** 10612.9200 − 30.5800 0.0000 − 345407.3000 − 303721.2000
STAGEit 376.6065*** 24.9264 15.1100 0.0000 327.6529 425.5602
DEPit − 0.2375 0.1405 − 1.6900 0.0920 − 0.5134 0.0385
SIZEit 10146.8500*** 1013.1080 10.0200 0.0000 8157.1720 12136.5200
RPMit 0.0000857*** 0.0000 14.1400 0.0000 0.0001 0.0001
ADAYSit − 0.3560 0.5431 − 0.6600 0.5120 − 1.4226 0.7106
LFit − 69408.8000 72534.2000 − 0.9600 0.3390 − 211861.1000 73043.4400
EUETS − 22421.6600 11808.7800 − 1.9000 0.0580 − 45613.3100 769.9903
AVGFAREit − 123.6948*** 30.8648 − 4.0100 0.0000 − 184.3111 − 63.0785
α − 1668174.0000 169750.0000 − 9.8300 0.0000 − 2001552.0000 − 1334797.0000
σu 234253.37
σe 68850.429
ρ 0.9204845 (fraction of variance due to ui )
R-sq:
within = 0.9738
between = 0.9959
overall = 0.8469
F test that all ui =0: F(2, 599) = 175.35
Prob > F = 0.0000
Notes: Legend: * p<0.05; ** p<0.01; *** p<0.001.
where an increased load factor means more passengers per flight and thus a lower amount of emission on per passenger basis. As our
study focuses on total carbon emissions, and the sector comprising of the three airlines combined, we are not investigating the effect of
these determinants on per passenger basis. Thus, now we can use these determinants along with technological innovation variables in
the simulation model to gauge the impact of CORSIA on the total carbon footprint of these airlines.
4. The simulation model
After identifying the critical determinants affecting carbon emissions for full-service carriers, we develop a system dynamics-based
simulation model to gauge the impact of CORSIA implementation on airline emissions. Herein, the econometric model’s results serve
as important inputs to the conceptual model. Further, we use the determinants along with biofuel adoption, new fleet inclusion and
CORSIA implementation to conceptualize the model.
Now, we first adopt the systems optimization technique, which focuses on dividing a big system into simpler sub-systems, while
optimizing them separately, which at times plagues the overall system due to loss of interconnected interactions among the sub-
systems. Then, we have the system dynamics modeling that is dynamically efficient. It incorporates feedbacks and interactions
among the sub-systems, time delays, accumulations and nonlinearities that are much closer to real world problems (Sterman, 2000).
Notably, the systems view avoids the twin dangers of mentality that follows the breaking down of the system into separate entities, viz.
(i) fixing a problem temporarily and (ii) organizational myopia where problems are dealt in a quick fix mode without thinking about its
occurrence for the worse form in future. Sherwood (2011) suggests the meaning of systems thinking as an approach to tame the complex
real world situations by looking at the system as a whole. Arnold and Wade (2015) defined systems thinking as a system of thinking
about the systems. Further, Meadows (2008) suggests that systems thinking comprises of three entities, namely, the elements, the
interconnections among them and an objective function or purpose. Thus, the system dynamics theory is well grounded in optimal
feedback control and non-linear dynamics developed in Mathematics, Science and Physics (Sterman, 2000), and thereby is apt for
dealing with complex design problems. Menon and Mahanty (2012) establish the suitability of the system dynamics approach in
energy as well as the transportation sector.
As a remark, in system dynamics modeling, two essential components of modelling are brought together to render an explicit
model; namely, the structure of the system (qualitative aspect), and its parameters (quantitative aspect) (Pfaffenbichler et al., 2010).
These two components form the complete simulation model, which we would discuss at length in the sub-sections ahead. Importantly,
the qualitative aspect is also known as the causal loop diagram (CLD) that discusses the conceptual building of a model; while the
quantitative aspect is also known as the stock flow diagram (SFD), that essentially brings the conceptual model to a simulation model
with the help of governing equations and parameters in order to analyze policy implications and scenarios that we develop.
12
4.1. The causal loop diagram: a four point CORSIA framework
When fully implemented, CORSIA would be a substantial and significant step towards global climate initiatives. As per Environ
ment Defense Fund (EDF) estimates, a fully implemented CORSIA could reduce around 2.5 billion tons of carbon emission during the
15 years of the policy time. We focus on four emission reduction ways suggested by the CORSIA policy framework to build our
conceptual model, using the determinants identified in the econometric model as significant influencers. The four ways stated in EDF
(2020) include (a) flying fuel efficient aircraft, (b) using new technologies to choose flight paths (network aspect), (c) using sustainable
alternative biofuels, and (d) investing in carbon offsets towards green projects to reduce carbon footprints. The resulting conceptual
model or the causal loop diagram (CLD) is shown in Fig. 3.
CLD is the conceptual model designed to capture the feedback structure of the modelled system (Sterman, 2010). CLD (Fig. 3)
consists of variables connected via arrows that explain the causal influences between the connected variables. Each causal link has an
appropriate polarity as per the nature of change experienced by the dependent variable, especially when there’s a change in the in
dependent variable. The (+) sign signifies that a dependent variable would move in the same direction as the independent variable,
while a (− ) sign suggests the change in dependent variable opposite to that of the independent variable. Some key loops are high
lighted using loop identifiers. A loop identifier suggests the nature of the feedback. A positive feedback is denoted by a reinforcing (R)
sign and a negative feedback is denoted using balancing (B) sign. In other words, when there is a net sign reversal, the disturbance
created by the independent variable is opposed. Thus, there’s a balance or check disturbance. However, when there’s no net sign
reversal, the disturbance amplifies, causing reinforcement thereof. The arrow surrounds the loop identifier suggests the direction of
flow of the loop. Please refer to Sterman (2010) for in-depth discussion of the system dynamics models. There are two key balancing
loops B1 and B2 and two reinforcing loops R1 and R2 capturing the essence of our study’s focus area that are explained to have an
insight into our conceptual model building exercise.
4.1.1. Balancing Loop B1
• FlightFrequency⟶Departures⟶DistanceFlown⟶TotalCO2Emission⟶CarbonOffsettingCharges⟶FlightFrequency
The causal influence of increased flight frequency would be the increased departures from the airports. Increased departures would
then lead to more distance being flown in total by airlines, resulting thereby in increased total carbon emissions. The higher the
emissions, the higher the offset would be. This entices the airline to surrender its earnings to green projects, while resulting in a check
on flight frequency via some alternate mechanism, without losing its revenues. Interestingly, the last relationship creates an opposition
to the increment of all the variables, bringing thereby a balance in the loop.
Fig. 3. The causal loop diagram.
13
4.1.2. Balancing Loop B2
• FlightFrequency⟶ DemandforAirTravel⟶ InternationalPassengers⟶ RevenuePassengerMiles⟶ TotalCO2 Emission⟶

CarbonOffsettingCharges⟶ FlightFrequency
A consistent flight frequency at regular intervals would attract more passengers towards air travel. Higher demand for air travel
over internal routes would increase the number of international passengers, increasing RPM thereby. Importantly, an increase in RPM
alludes to an exponential increase in the total carbon emissions. Thus, the higher the emissions, the higher would be the airlines’ offset.
The airline hence needs to surrender its earnings to green projects, while resulting in a check on flight frequency.
4.1.3. Reinforcing Loop R1
• ProfitMargin⟶ Investment⟶ NewAircraftFleet⟶ FuelEfficiency⟶ FuelConsumption⟶ TotalCO2 Emission⟶

CarbonOffsettingCharges⟶ ProfitMargin
An increase in profit margins would increase investment opportunities in newer fuel-efficient fleets. The higher the proportion of
new fleet, the higher would be the fuel efficiency of overall fleet, and consequently lesser would be the fuel consumption. Nevertheless,
higher fuel consumption implies higher total carbon emission level, resulting in higher offsets for the airline. Finally, the higher the
carbon offsetting requirements, the lesser the profit margins would be. Notably herein, opposition to the disturbance happens twice,
thus the net reversal in disturbance has to be negated in order to trigger the reinforcement effect.
4.1.4. Reinforcing Loop R2
• ProfitMargin⟶Investment⟶AlternativeBioFuels⟶TotalCO2Emission⟶CarbonOffsettingCharges⟶ProfitMargin
An increase in profit margins would increase the investment opportunities in technologically advanced alternative fuels which emit
less carbon in the environment. An increase in the use of biofuels would result in a decrease in total carbon emissions. The higher the
emissions, the higher the offsets would be. Thus, more the airline has to surrender its earnings to green projects, resulting in lesser
profit margins. There are some other minor loops in the conceptual model that help to estimate the effect on some variables. This
conceptual model covers the targeted aspects to be studied as per the four ways discussed earlier to reduce emission under CORSIA
implementation. CLD brings us to the structure of the system and make the modeled system ready for simulation.
4.2. The stock flow diagram
The stocks and flows are the central components of the system dynamics models. Stocks are the accumulations in a system; they are
the entities that give inertia to the modeled system, serving as the system’s memory. Furthermore, they are the level variables within a
model that create delays by accumulating the net change induced from both inflows and outflows. Flows, on the other hand, are the
auxiliary variables that have a rate of change. In general, flows are the functions of stock and other variables in the system. The final
stock flow diagram for the simulation runs is shown in Fig. 4. The mathematical expressions explaining the stocks and rate of change of
Fig. 4. The stock flow diagram.
14
stocks are shown by Eq. (6) and Eq. (7), respectively (Sterman, 2010).
∫t
Stock(t) = [Inflow(s) − Outflow(s)]ds + Stock(t0 ), (6)
t0
d(Stock)/dt = Inflow(t) − Outflow(t). (7)
4.2.1. Stocks studied under the modeled system

There are two major stocks that are studied under the modeled system. The first represents the total carbon emission from the
beginning of simulation duration till the policy completion year (i.e. 2035). The total carbon the three airlines would have emitted in
the environment would be measured by this stock. The second stock is the total carbon offsetting being done by the airlines during the
policy period, starting from 2021 till the policy completion year (i.e. 2035). The mathematical expression of the two stocks based on
standard equations defined in Sterman (2010)’s study, are shown in Eq. (8) and Eq. (9) respectively. Table 7 gives the details about the
variables, parameters, and constants used in the model along with their units and values.
∫t
TotalCO2 (t) = [EmissionEnablers(s) − EmissionInhibitors(s)]ds + TotalCO2 (t0 ), (8)
t0
∫ t
CarbonOffsetting(t) = [OffsettingRequirement(s) − CORSIAExemptions(s)]ds + CarbonOffsetting(t0 ). (9)
t0
Table 7
Entities in the simulation model and their description.
Entity Description Type Value Range Unit
Conventional Fuel Fraction Fraction of conventional fuel used out of total consumption auxiliary variable equation driven dimensionless
Distance Flown Total internatonal distance flown by the airlines auxiliary variable actual data km per Month
Baseline Fuel Consumption Fuel consumption in current situation auxiliary variable actual data kg per Month
Total Fuel Consumption Fuel consumption based on distance flown auxiliary variable actual data kg per Month
Emission From Fuel Burn Total emission generated from fuel burn auxiliary variable equation driven kg per Month
CO2 Footprint Reduction Carbon footprint reduction using technology auxiliary variable equation driven Tonnes per Month
Emission Enablers Total emission generated from all sources considered auxiliary variable equation driven Dollars per Month
Emission Inhibitors Total emission mitigated from all sources considered auxiliary variable equation driven Dollars per Month
Monthly Carbon Emission Net CO2 emission by airlines auxiliary variable equation driven Dollars per Month
Offsetting Requirement Amount of emission required to be offset auxiliary variable equation driven Dollars per Month
CORSIA Exemptions Amount of emission exempted from the offset auxiliary variable equation driven Dollars per Month
AA American Airlines binary variable {0,1} dimensionless
DA Delta Airlines binary variable {0,1} dimensionless
UA United Airlines binary variable {0,1} dimensionless
BFC Baseline fuel consumption switch binary variable {0,1} dimensionless
ABF Alternative biofuel switch binary variable {0,1} dimensionless
Implementation Policy implementation switch binary variable {0,1} dimensionless
New Fleet Fuel efficient fleet induction switch binary variable {0,1} dimensionless
CFER kg CO2 per kg of conventional jet fuel burnt constant 3.16 dimensionless
ABFER kg CO2 per kg of alternative biofuel burnt constant {0.158, 0.79, 1.58} dimensionless
TF Tonne Factor constant 0.001 Tonne per kg
Exemption Limit Tonnes of CO2 exempted from offsets constant 10000 Tonnes
Medium Haul Airline Fraction Fraction of medium haul flights on international routes constant (0,1) dimensionless
Medium Haul Aircraft FBR Fuel burn rate of medium haul aircraft constant {2.91, 2.99, 3.79} kg per km
Long Haul Aircraft FBR Fuel burn rate of long haul aircraft constant {5.94, 7.18, 8.58} kg per km
Total CO2 Emission $ Total net CO2 emission over the period level variable equation driven Dollars
Carbon Offsetting $ Total net carbon offsetting over the period level variable equation driven Dollars
Fuel Efficiency of New Fleet Fraction by which new fleet is efficient over the current fleet lookup (0,1) dimensionless
Fuel Efficient Fleet Fraction Fraction of new fleet out of total fleet lookup (0,1) dimensionless
CORSIA CORSIA’s level of implementation lookup {0,1} dimensionless
IGF Individual growth factor for y-o-y emission lookup (0,1) dimensionless
SGF Sector growth factor for y-o-y emission lookup (0,1) dimensionless
SGF Fraction fraction of SGF to be used in calculating offsets lookup (0,1) dimensionless
FCAA Fuel Consumption by American Airlines lookup actual data kg per Month
FCDA Fuel Consumption by Delta Airlines lookup actual data kg per Month
FCUA Fuel Consumption by United Airlines lookup actual data kg per Month
DFAA Distance flown by American Airlines lookup actual data km per Month
DFDA Distance flown by Delta Airlines lookup actual data km per Month
DFUA Distance flown by United Airlines lookup actual data km per Month
Alternative Bio Fuel Fraction Fraction of alternative biofuel used out of total consumption lookup (0,1) dimensionless
15
4.2.2. Simulation setup and scenarios

The simulation model developed aims at gauging the carbon emission behavior of the airlines under consideration. The year from
where the observation starts is from the first month of 2003, and final year of observation is 2035 till the last month. The total analysis
period in the simulation is 32 years, with 2021 being the beginning of the 15 years of CORSIA implementation period. The unit of time
is month. There are four major scenarios that have been developed to analyze the impact of different components of technologies and
policies, adopted to reduce the total emissions by the full-service carriers. All the binary variables described in Table 7 act as a switch to
either enable or disable an attribute under consideration. The first scenario is the Status Quo, where no technological or policy in
terventions are taken into account. The next three scenarios differ in the degree and extent of adoption of different means to reduce
carbon footprints. The Pessimistic Scenario describes a situation, where the attributes are adopted in next 15 years to their minimum
levels, wherein CORSIA is not adopted. The bio-fuel used in this scenario is assumed to have the minimum carbon reduction efficiency
of 50% (IRENA, 2017) based on vegetable oil extracted from feed stocks.
The aircraft used by the airlines are the ones that are based on their actual fleet configuration. In this scenario, it is assumed that
only re-engineering would be done to improve the fuel efficiency to a maximum of 15%. The Optimistic Scenario, on the other hand,
assumes the maximum adoption of all the attributes by the airlines, including CORSIA implementation. The maximum prospective
market for bio-fuels i.e. 50%, (IRENA, 2017) is adopted in this scenario. The aircraft fleet is expected to be upgraded to a range of
minimum 60% to a maximum of 80% by the end of 2035. Further, between these two extremes is the Realistic Scenario where CORSIA is
adopted, and the attributes are assumed to fall at a middle ground between the other two scenarios. Here, it is assumed that at the
maximum 60% of the older fuel inefficient fleet would be replaced, and at least 30% of conventional jet fuel would be replaced by bio-
fuels. The summary of all the scenarios is tabulated in Table 8. Note that, all the simulation runs are carried out in Vensim from
Ventana Systems, and graphical plots are generated for different stocks and flows.
5. Results
We conduct a number of simulation runs considering various aspects described in the four scenarios, and an additional set of runs to
gauge the impact of each component of emission reduction when adopted by the airlines. All these experiments aim at analyzing the
changes in carbon footprints and offsetting requirements when CORSIA is in place and otherwise. Notably, all the four scenarios along
with the simulation runs have been performed on a workstation with Intel Core i7-2600 3.40 GHz CPU, and 32 GB of RAM.
Table 8
Simulation Scenarios.
Component Attribute Status Quo Scenario Pessimistic Scenario Realistic Scenario Optimistic Scenario
Emission Measure Approach

Current ✓ - - -
{95% of (2004–2006) levels}
Distance based - ✓ ✓ ✓
{econometric measure}
Fuel Usage
Alternative Biofuel - ✓ ✓ ✓
Alternative Biofuel Fraction 0.00 0.2–0.3 0.3–0.4 0.4–0.5
ABFER - 1.580 0.790 0.158
Conventional Jet Fuel ✓ ✓ ✓ ✓
Conventional Jet Fuel fraction 1.00 0.7–0.8 0.6–0.7 0.5–0.6
CFER 3.16 3.16 3.16 3.16
Technologically advanced Fleet
Fuel Efficient Fleet - ✓ ✓ ✓
Fuel Efficiency - 0.15 0.18–0.22 0.22–0.43
Fraction of New Fleet - 0.2–0.4 0.4–0.6 0.6–0.8
Fleet Configuration
Medium Haul Flights current ✓ ✓ ✓
Aircraft - Boeing 757 Boeing 737–800 Airbus A321NeoLR
FBR - 3.79 2.91 2.99
Fraction of total flights - 0.35 0.60 0.80
Long Haul Flights current ✓ ✓ ✓
Aircraft - Boeing 777 Boeing 787 Airbus A330Neo
FBR - 8.58 7.81 5.94
Fraction of total flights - 0.65 0.40 0.20
Policy Implementation
CORSIA - - ✓ ✓
Notes: All the data related to biofuels and the maximum biofuel adoption of 50% (market projections) has been taken from IRENA (2017). The fuel
efficiency estimated are based on new aircraft’s entry into service (EIS) discussed in Ploetner et al. (2017). The maximum value of each fuel efficiency
range are for the years 2030–2035.
16
5.1. Impact of each component on carbon footprints
Before undergoing simulation runs for the different scenarios, we first analyze the impact of each component, shown in Table 8, on
carbon footprints. The base case is the adoption of distance based carbon emission evaluation (TSF), followed by adoption of alter
native biofuels (TSF + ABF), fuel efficient aircraft fleet induction (TSF + ABF + FEF) and finally the CORSIA policy implementation
(TSF + ABF + FEF + COR). The results are shown in Fig. 5. Fig. 5 (a) shows the total carbon emission on a monthly basis by the three
airlines combined. Herein, it must be noted that as each component is adopted by airlines, the carbon emission levels reduce with the
largest reductions observed when CORSIA is in place, followed by the case when alternative biofuels are adopted. Fig. 5 (b) captures
the total carbon emission levels over the analysis period. It is clear from the figure that the trajectory only starts to flatten when an
offsetting scheme is put in place. Further analysis under each scenario would suggest the achievable gains by incorporating each
attribute, and how each component has been contributing towards a carbon neutral growth. Moreover, some route related insights are
also expected by fleet configuration component.
5.2. Main simulation results
5.2.1. Aircraft types, the routes and the fuel consumption pattern
Fig. 6 (a) elaborates upon the total fuel consumption patterns by the airlines under different scenarios. The Status Quo scenario
doesn’t use the distance-based emission approach; thus, fuel consumption for this case is not shown in Fig. 6 (a). However, when the
airlines keep using the current fleet with re-engineering, the total fuel consumption is at the maximum level. With the adoption of
newer aircrafts like Airbus A321Neo LR and Airbus A330Neo, the airlines substantially reduce their fuel consumption. Herein, it must
be noted that the reduction in fuel consumption also stems from the proportion of medium haul to long haul flights. With increasing
medium haul flights, the airlines tend to use aircraft with lower fuel burn rate than the aircraft used in the long haul flights, reducing
thereby fuel consumption. The more medium haul routes are adopted in the international sky, the lesser is the fuel consumption by the
airlines over the same total distance flown, as the aircraft like Airbus 321Neo LR or Boeing 737–800 are more fuel efficient.
5.2.2. Total emission from fuel burn and carbon footprint reduction
It is clear from Fig. 6 (b) that the total emission emanating from the Status Quo scenario is maximum, as no component of emission
reduction is adopted in this case. Further, as the fuel consumption pattern suggests, the more the airlines burn fuel, the more their
carbon emission is. Even with the attribute levels adopted under the Pessimistic scenario, emissions originating from fuel consumption
reduces drastically in comparison to the levels when status quo is maintained. With increasing percentage of biofuels and fuel-efficient
aircraft fleet, emission from fuel burn takes a further plunge under both Realistic and Optimistic scenarios. 6 (c) shows the reducing
fraction of conventional jet fuel over the years across different scenarios. From the Fig. 6 (d), the effect of advanced aircraft fleet is
predominantly visible. As the percentage of fuel-efficient aircraft increases in the fleet, the carbon footprint decreases. There is a
certain improvement in the reduction from 2030 onward across the three scenarios, as the new fleets have commissioned to use the
higher fuel efficiency aircraft. These amplifications in the carbon footprint reduction are visible in the net carbon emission levels.
5.2.3. Total emission levels and offsetting requirements

In order to limit the growth of carbon emission in the environment, it is certain that maintaining the status quo would help in no
way. Fig. 7 (a) shows the monthly carbon emission levels under the four scenarios. Even if the goal is just to maintain the current
emission levels for the next 15 years, we need to adopt the cleaning attributes, at least under the Pessimistic scenario. Although, the real
impact and reduction would come only when CORSIA is in place, be it the Realistic or the Optimistic scenario. Further, the total carbon
emission levels flatten under both the latter scenarios as shown in 7 (b). Without carbon offsetting scheme, the total carbon emission is
going to increase consistently under both Status Quo and Pessimistic scenarios. The carbon offsetting requirements when CORSIA is
Fig. 5. Impact analysis of different components of CO2 reduction initiatives.
17
Fig. 6. Simulation results on fuel consumption and emission patterns.
implemented is shown in 7 (c) and the total offsets are shown in 7 (d). For the first two scenarios, no policy is implemented, so the
figures contain two curves only. Moreover, the extent of adoption of biofuels, along with fuel-efficient fleets, varies among both
Realistic and Optimistic scenarios. Thus, the amount of carbon offsetting requirements are much higher, especially in the former sce
nario, as the emission levels are also higher. The total offsets by the airlines have nonlinear growth in both the scenarios. These offsets
result in green projects, which in turn reduces the total carbon emission substantially as evident in 7 (b).
6. Model validation
Systems dynamics modelers have addressed the issue of diagnosing flaws and errors in the models by describing a number of tests to
check and validate the models. To ensure research rigor, a few important model validation exercises that we performed to ensure the
consistency of results are as follows.
6.1. Dimensional consistency check
In order to have the correct governing equations among various interacting variables, dimensional consistency runs are performed
in the Vensim unit check module. The units of each variable and parameters are defined and tabulated in Table 7. Notably, the stock
and flow model is dimensionally consistent and it fits for the simulation runs.
6.2. Extreme conditions test
Sterman (2010) emphasizes on having the extreme conditions test, which uncovers whether each governing equation makes sense
under extreme conditions. All the switches (a 0–1 switch to implement a policy) defined in Table 7 are results of multiple iterations of
the model but under extreme conditions and values. Each time a model performs unexpectedly, it is identified that a switch is needed to
regulate correctly the conceptualized idea. For example, when the Implementation switch is set to an extreme condition of 0, it signifies
that CORSIA is not implemented. Thus, none of the impacts of the policy, such as offsetting requirements or emission inhibited due to
green projects exist in the model analysis. A model is robust only when each of these aspects are ensured. Thus, extreme condition test
ensures a robust model response. Once all the switches are in place, the model behaves sensibly at all the extreme values, ensuring
plausibility at extreme policy, parameters and shocks.
18
Fig. 7. Simulation results on total emissions and offsetting.
6.3. Statistical validation of simulation results
Summary statistics of historical fit are standard validity tests adopted for systems dynamics simulation models (Oliva, 1995). This
statistical fit test has been used since 1980s (Sterman, 1984). The summary statistics show a relation between both the historical and
the simulated data for all the variables under study. We performed the statistical fit test for two of the major variables in our simulation
model, i.e. fuel consumption pattern and the CO2 emission pattern. A lot of equations are driven by these variables; thus, the pattern
obtained for these variables through simulation run are tested statistically with the historical data available. A summary statistics for
historical fit is tabulated in Table 9.
Furthermore, Table 9 suggests a high correlation coefficient between the simulated and the historical values of the variables under
study. This instills confidence in the model and further ensures that the model is statistically significant, and in congruence with the
Table 9
Summary statistics for historical fit.
Statistics Simulated v/s Historical Simulated v/s Historical
Fuel Consumption CO2 Emission
Correlation Coefficient 0.8348 0.5638

R Square 0.6969 0.3179
Adjusted R Square 0.6767 0.2906
Coefficients 2.1780 3.9041

Standard Error 0.3708 1.1437
t-Stat 5.8731 3.4136
P-value 0.0000 0.0021
Lower 95% 1.3876 1.5486
Upper 95% 2.9685 6.2596
Notes: Fuel consumption is significant at p-value < 0.001 while CO2 Emission at p-value < 0.01. Coyle and Exelby (2000) suggest that there is no such
thing as getting an absolute validity for the models, instead the confidence gets improved when validation tests are performed.
19
actual data. The plot of statistical fit between simulated and historical fuel consumption and CO2 emission are shown in Fig. 8.
7. Managerial implications
There are some clear policy implications and win–win situation for both the aviation sector and the environment that are evident
from various analyses we have performed. They would help policy makers and managers in multiple ways.
7.1. The determinants of carbon emissions
The econometric analysis signifies the importance of knowing the proper determinants of carbon emission for the airlines. It helps
not only in understanding the factors that are responsible for carbon emission, but also serves in designing a proper emission offsetting
scheme. The average stage length, being a positive determinant of carbon emission, suggests that distance is an important measure of
calculating carbon emissions per flight basis. The distance-based approach to calculate carbon emissions opens up new perspectives
about the planning on how to reduce the air time per flight in international skies, especially when policies like CORSIA are in place. The
load factor, does not come out to be a significant determinant of total airline emission. However, for calculating the carbon emissions
per passenger, it plays an important role. Thus, managers focusing on per passenger emissions should look at the load factor from
productivity enhancement perspectives. Aircraft days being an insignificant determinant of total emissions, supported by Dray (2013),
signifies that total emissions depend on more directly impacting factors like flight frequency, stage length, fuel consumption and
aircraft size. Further, the identified determinants would help both managers and planners to determine better models to predict carbon
emissions.
7.2. Aircraft size, stage length and route configuration
As evident from the econometric analysis, the size of the aircraft and the average stage length play a significant role in the total
carbon emission that an airline generates. Larger aircraft are not desirable from an emission perspective. Givoni and Rietveld (2009)
substantiate these findings in their study, and suggest that a growing demand of air travel is better catered by increasing flight fre
quency and not by aircraft size. Further, it is important for airlines to optimize the configuration of aircraft in the fleet. A dispro
portionate amount of larger aircraft in the fleet would mean a higher fuel burn rate per unit distance. This would thereby compel the
airline to fly longer distances in order to gain economy. Importantly, both these factors would increase the carbon footprints of the
airlines. This concern is further supported by our simulation results, by varying the proportion of medium to long haul flights in the
fleet. The higher the number of medium haul aircraft, the lesser are the fuel burn rate and total resultant emissions in the international
skies. Our suggestion of higher medium haul flights and more frequent flights is supported by Givoni and Rietveld (2010) and Sibdari
et al. (2018). Airlines must reconsider their route configurations to reduce their average stage length, their air time per flight in the
international skies in order to reduce international emissions, resulting in optimized carbon offsets to be surrendered. Thus, CORSIA
would encourage airlines to reconfigure their international route configurations, and their medium to long haul fleet proportions as
well, supporting the need for shorter international routes.
7.3. Alternative biofuels and fuel efficient aircraft fleets
CORSIA alone cannot do the job of cleaning the environment we breathe in. Indeed, it must be supported by technological
Fig. 8. Simulation results on total emissions and offsetting.
20
innovations in generating cleaner fuels and fuel efficient fleets. Alternative biofuels may take time to gain economic viability as well as
the volumes required to replace jet fuels, but the emission reducing efficiencies obtainable from biofuels would surely reduce a sig
nificant amount of GHGs from the environment. A systematic vision and commitment towards the adoption of these environment-
friendly fuels would help us achieve CNG (2020). Moreover, innovations in designing fuel efficient engines would help us conserve
fuel, and thereby reduce the level of emissions per flight. Our study addresses one of the limitations of Lo et al. (2018), where due to
lack of data, they have ceased to compare the emission by aircraft type. However, we have compared the performance of different
aircraft under pessimistic, realistic and optimistic scenarios. We also compared medium haul to long haul flights, and suggested a usage
combination in order to curtail carbon footprints. Further, our study has used actual distance flown by an aircraft, and not the
theoretical ones.
7.4. CORSIA implementation and participation
CORSIA is no doubt an initiative that is timely and critical. Airlines can no longer continue to emit higher levels of GHGs to pollute
the environment when frequent negative impacts of global warming, melting glaciers and recurring wildfires are evident. Making an
airline responsible for offsetting the additional emission beyond the allowable limits would help environment in an affirmative way,
either through investments in green projects or through reduction through technological advancements. It would compel the airlines to
opt for emission reduction technologies, while restructuring their flight operations in the international skies. The goals of carbon
neutral growth would be achievable only when airlines commit towards a responsible operation of the sector of transportation that
gives human beings the wings to fly.
7.5. Regional Impact of CORSIA in different parts of the world
The fact that CORSIA is a global offsetting mechanism makes it a policy that would affect the international civil aviation in all parts
of the world. Additionally, as per Assembly Resolution A39-3 and Chicago Convention, ICAO member states have to comply to
Standards and Recommended Practices (SARPs) in their national aviation policies. Each member state has to ensure that her domestic
policies are aligned with these SARPs. The aim of this uniformity is to keep the markets away from distortions while protecting the
environmental integrity of CORSIA. Thus, along with international emissions regulation, CORSIA will indirectly impact and
restructure the domestic policies of the states.
Developing states that have their own capabilities and concerns are allowed not to be a part of CORSIA voluntarily from the
beginning till 2027 and thus the policy is rolled out in a phased manner. However, non-participating states also have to monitor their
emission levels and report the same. There are some exceptions where mandatory participation would be waved off. For example, the
states emitting less than 0.5% of international airline emissions, the least developed states, landlocked or very small island states and
states in primitive phases of development would be exempted from mandatory offsetting.
7.6. Role of airlines and the airports in CORSIA regime
One of the crucial areas of attention for the airlines is the aviation fuel. Currently, the conventional fuels lead in all fronts, be it the
established infrastructure, efficiency of the distribution systems, or the profitability. One thing that downplays the entire conventional
fuel segment is the green house emission that happens as a consequence of using them. Most air journeys emit more green house
emissions than their road and rail counterparts. Thus, in order to curb the emission levels and to achieve the CNG-2020 goals, airlines
have to (i) adopt alternative biofuels, (ii) extensively promote voluntary carbon offsetting and (iii) encourage the passengers to be a
part of the whole process. Some airlines are already doing it, like QANTAS. Further, airlines have to plan their fleet renewal and
adoption of new generation aircraft that emit much less than the current generation of aircraft.
Similarly, airport authorities must participate in upgrading the infrastructure, promote design of green airports and provide
necessary infrastructure for alternative biofuel storage and distribution. Moreover, airports can provide better slots to green airlines
over polluting airlines in order to encourage the airlines to upgrade themselves. These initiatives are not going to happen overnight,
but a gradual move towards these changes would decide the future of emission free skies.
8. Conclusion and future work
The study explored important questions regarding carbon footprints left by the aviation sector, its determinants, and the impact of
CORSIA implementation in international air travel. We developed an econometric model to identify and investigate the determinants
of carbon emission for airlines. We identified the significant determinants with their magnitude and direction of affecting the total
carbon emissions. The econometric model utilized panel data of 17 years for the three largest full-service carriers in the US aviation
sector. Further, the model used the fixed effect panel data regression technique to investigate the hypotheses. Our results indicated that
fuel cost per gallon, the average airfare, EU ETS and the total departures are negative determinants of emissions. Whereas, average
stage length, aircraft size and revenue passenger mile have a positive relationship with total emissions. The aircraft days and load
factor show a positive but insignificant relationship with total emissions. Based on the significant determinants along with the major
technological innovation-based factors, we designed a simulation model to gauge the impact of each attribute on the total carbon
emissions of the airlines when CORSIA is implemented. We used Vensim to simulate the various scenarios differing in their extent of
adoption of carbon mitigation measures.
21
The analyses done in the study examines the CORSIA implementation which is a global offsetting mechanism. Only with the ex
ceptions of firefighting events, humanitarian crisis and medical flights, all the international civilian airline operations will fall under
CORSIA including training and maintenance flights, all schedule and non-schedule flight operations and passenger as well as cargo
flights operating in international skies. Although, the analyses carried out in the current study uses data from the full-service carriers in
the USA, the implications and impact of CORSIA is global in all parts of the world. In other words, as long as a flight operates in an
international sky, it will fall under CORSIA regulations as it is the first unified international civil aviation policy. The study suggested
various courses of action to both decision-makers and managers, which would affect the airlines as well as the environment. However,
there are regions where further improvements may be explored. One such limitation is exploring from the customer’s perspective over
the possibility of increased airfares when airlines would have carbon offsets to surrender. Future works could explore the customer’s
willingness to share emission related burden, as well as a mechanism in which it would be viable for airlines to engage the customers
rather than imposing straightforward airfare increments on their pockets. Future study could also explore a mechanism by analyzing
real-life scenarios relevant to the managers and the policy-makers.
References
Albers, S., Bühne, J.-A., Peters, H., 2009. Will the eu-ets instigate airline network reconfigurations? J. Air Transport Manage. 15 (1), 1–6.
Ambrose, J., 2020. Airbus reveals plans for zero-emission aircraft fuelled by hydrogen. The Guardian.
Anger, A., 2010. Including aviation in the european emissions trading scheme: Impacts on the industry, co2 emissions and macroeconomic activity in the eu. J. Air
Transport Manage. 16 (2), 100–105.
Ansell, P.J., Haran, K.S., 2020. Electrified airplanes: A path to zero-emission air travel. IEEE Electrification Mag. 8 (2), 18–26.
Arnold, R.D., Wade, J.P., 2015. A definition of systems thinking: A systems approach. Procedia Comput. Sci. 44 (2015), 669–678.
Arul, S.G., 2014. Methodologies to monetize the variations in load factor and ghg emissions per passenger-mile of airlines. Transport. Res. Part D: Transport Environ.
32, 411–420.
Belobaba, P., Odoni, A., Barnhart, C., 2015. The Global Airline Industry. John Wiley & Sons.
Brueckner, J.K., Abreu, C., 2017. Airline fuel usage and carbon emissions: Determining factors. J. Air Transport Manage. 62, 10–17.
Brueckner, J.K., Zhang, A., 2010. Airline emission charges: Effects on airfares, service quality, and aircraft design. Transport. Res. Part B: Methodol. 44 (8–9),
960–971.
Brugnoli, A., Button, K., Martini, G., Scotti, D., 2015. Economic factors affecting the registration of lower co2 emitting aircraft in europe. Transport. Res. Part D:
Transport Environ. 38, 117–124.
Carlsson, F., Hammar, H., 2002. Incentive-based regulation of co2 emissions from international aviation. J. Air Transport Manage. 8 (6), 365–372.
Chao, C.-C., 2014. Assessment of carbon emission costs for air cargo transportation. Transport. Res. Part D: Transport Environ. 33, 186–195.
Chao, H., Agusdinata, D.B., DeLaurentis, D., Stechel, E.B., 2019. Carbon offsetting and reduction scheme with sustainable aviation fuel options: Fleet-level carbon
emissions impacts for us airlines. Transport. Res. Part D: Transport Environ. 75, 42–56.
Coyle, G., Exelby, D., 2000. The validation of commercial system dynamics models. Syst. Dyn. Rev.: J. Syst. Dyn. Soc. 16 (1), 27–41.
Cui, Q., 2019. The online pricing strategy of low-cost carriers when carbon tax and competition are considered. Transport. Res. Part A: Policy Practice 121, 420–432.
Cui, Q., Li, Y., 2017. Airline efficiency measures under cng2020 strategy: An application of a dynamic by-production model. Transport. Res. Part A: Policy Practice
106, 130–143.
Cui, Q., Li, Y., Lin, J.-L., 2018. Pollution abatement costs change decomposition for airlines: An analysis from a dynamic perspective. Transport. Res. Part A: Policy
Practice 111, 96–107.
EDF, 2020. Reducing aviation’s climate impact.
Evans, A.D., 2014. Comparing the impact of future airline network change on emissions in india and the united states. Transport. Res. Part D: Transport Environ. 32,
373–386.
Fukui, H., Miyoshi, C., 2017. The impact of aviation fuel tax on fuel consumption and carbon emissions: The case of the us airline industry. Transport. Res. Part D:
Transport Environ. 50, 234–253.
Givoni, M., Rietveld, P., 2009. Airline’s choice of aircraft size–explanations and implications. Transport. Res. Part A: Policy Practice 43 (5), 500–510.
Givoni, M., Rietveld, P., 2010. The environmental implications of airlines’ choice of aircraft size. J. Air Transport Manage. 16 (3), 159–167.
Grampella, M., Lo, P.L., Martini, G., Scotti, D., 2017. The impact of technology progress on aviation noise and emissions. Transport. Res. Part A: Policy Practice 103,
525–540.
Hausman, J.A., 1978. Specification tests in econometrics. Econometrica: J. Eonom. Soc., 1251–1271.
Hofer, C., Dresner, M.E., Windle, R.J., 2010. The environmental effects of airline carbon emissions taxation in the us. Transport. Res. Part D: Transport Environ. 15 (1),
37–45.
ICAO, 2017. Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA).
IRENA, 2017. Biofuels for aviation: Technology brief.
Liu, X., Hang, Y., Wang, Q., Zhou, D., 2020. Drivers of civil aviation carbon emission change: A two-stage efficiency-oriented decomposition approach. Transport. Res.
Part D: Transport Environ. 89, 102612.
Liu, X., Zhou, D., Zhou, P., Wang, Q., 2017a. Dynamic carbon emission performance of chinese airlines: a global malmquist index analysis. J. Air Transport Manage.
65, 99–109.
Liu, X., Zhou, D., Zhou, P., Wang, Q., 2017b. What drives co2 emissions from china’s civil aviation? an exploration using a new generalized pda method. Transport.
Res. Part A: Policy Practice 99, 30–45.
Lu, J.-L., Shon, Z.Y., 2012. Exploring airline passengers’ willingness to pay for carbon offsets. Transport. Res. Part D: Transport Environ. 17 (2), 124–128.
Maertens, S., Grimme, W., Scheelhaase, J., 2020. Icao’s new corsia scheme at a glance–a milestone towards greener aviation. Aviation and Climate Change: Economic
Perspectives on Greenhouse Gas Reduction Policies.
Meadows, D.H., 2008. Thinking in systems: A primer. chelsea green publishing.
Menon, B.G., Mahanty, B., 2012. Effects of fuel efficiency improvements in personal transportation. Int. J. Energy Sect. Manage.
Merkert, R., Hensher, D.A., 2011. The impact of strategic management and fleet planning on airline efficiency–a random effects tobit model based on dea efficiency
scores. Transport. Res. Part A: Policy Practice 45 (7), 686–695.
Miranda, V.A., Oliveira, A.V., 2018. Airport slots and the internalization of congestion by airlines: An empirical model of integrated flight disruption management in
brazil. Transport. Res. Part A: Policy Practice 116, 201–219.
Miyoshi, C., Fukui, H., 2018. Measuring the rebound effects in air transport: The impact of jet fuel prices and air carriers’ fuel efficiency improvement of the european
airlines. Transport. Res. Part A: Policy Practice 112, 71–84.
Miyoshi, C., Mason, K.J., 2009. The carbon emissions of selected airlines and aircraft types in three geographic markets. J. Air Transport Manage. 15 (3), 138–147.
Morrell, P., 2009. The potential for european aviation co2 emissions reduction through the use of larger jet aircraft. J. Air Transport Manage. 15 (4), 151–157.
Nava, C.R., Meleo, L., Cassetta, E., Morelli, G., 2018. The impact of the eu-ets on the aviation sector: Competitive effects of abatement efforts by airlines. Transport.
Res. Part A: Policy Practice 113, 20–34.
22
Odeck, J., 2019. Estimating and predicting the operational costs of road tolls: An econometric assessment using panel data. Transport. Res. Part A: Policy Practice 130,
466–478.
Oliva, R., 1995. A vensim module to calculate summary statistics for historical fit. d-4584. System Dynamics Group, MIT, Cambridge, MA.
Pfaffenbichler, P., Emberger, G., Shepherd, S., 2010. A system dynamics approach to land use transport interaction modelling: the strategic model mars and its
application. Syst. Dyn. Rev. 26 (3), 262–282.
Ploetner, K.O., Rothfeld, R., Urban, M., Hornung, M., Tay, G., Oguntona, O., 2017. Technological and operational scenarios on aircraft fleet-level towards atag and
iata 2050 emission targets. In: 17th AIAA Aviation Technology, Integration, and Operations Conference, p. 3771.
Preston, H., Lee, D.S., Hooper, P.D., 2012. The inclusion of the aviation sector within the european union’s emissions trading scheme: What are the prospects for a
more sustainable aviation industry? Environ. Develop. 2, 48–56.
Qiu, R., Xu, J., Xie, H., Zeng, Z., Lv, C., 2020. Carbon tax incentive policy towards air passenger transport carbon emissions reduction. Transport. Res. Part D:
Transport Environ. 85, 102441.
Reynolds-Feighan, A., 2018. Us feeder airlines: Industry structure, networks and performance. Transport. Res. Part A: Policy Practice 117, 142–157.
Rotaris, L., Giansoldati, M., Scorrano, M., 2020. Are air travellers willing to pay for reducing or offsetting carbon emissions? evidence from italy. Transport. Res. Part
A: Policy Practice 142, 71–84.
Scheelhaase, J., Grimme, W., Schaefer, M., 2010. The inclusion of aviation into the eu emission trading scheme–impacts on competition between european and non-
european network airlines. Transport. Res. Part D: Transport Environ. 15 (1), 14–25.
Scheelhaase, J., Maertens, S., Grimme, W., Jung, M., 2018. Eu ets versus corsia–a critical assessment of two approaches to limit air transport’s co2 emissions by
market-based measures. J. Air Transport Manage. 67, 55–62.
Sherwood, D., 2011. Seeing the forest for the trees: a manager’s guide to applying systems thinking. Nicholas Brealey Int.
Sibdari, S., Mohammadian, I., Pyke, D.F., 2018. On the impact of jet fuel cost on airlines’ capacity choice: Evidence from the us domestic markets. Transport. Res. Part
E: Logistics Transport. Rev. 111, 1–17.
Sterman, J., 2010. Business dynamics. Irwin/McGraw-Hill c2000.
Sterman, J.D., 1984. Appropriate summary statistics for evaluating the historical fit of system dynamics models. Dynamica 10 (2), 51–66.
Toke, D., 2008. Trading schemes, risks, and costs: the cases of the european union emissions trading scheme and the renewables obligation. Environ. Plann. C:
Government Policy 26 (5), 938–953.
Torres-Reyna, O., 2007. Panel data analysis fixed and random effects using stata (v. 4.2). Data & Statistical Services, Priceton University, pp. 1–40.
Vespermann, J., Wittmer, A., 2011. Financial, ecological and managerial impacts of emission trading schemes: the case of lufthansa. Business Strategy Environ. 20 (3),
174–191.
Winchester, N., McConnachie, D., Wollersheim, C., Waitz, I.A., 2013. Economic and emissions impacts of renewable fuel goals for aviation in the us. Transport. Res.
Part A: Policy Practice 58, 116–128.
Yamaguchi, K., 2010. Voluntary co2 emissions reduction scheme: analysis of airline voluntary plan in japan. Transport. Res. Part D: Transport Environ. 15 (1), 46–50.
Yuen, A.C., Zhang, A., 2011. Unilateral regulation and greenhouse gas emissions: the case of the airline industry. Transport. Res. Part D: Transport Environ. 16 (7),
540–546.
Zou, B., Elke, M., Hansen, M., 2012. Evaluating air carrier fuel efficiency and co2 emissions in the us airline industry. The National Center of Excellence for Aviation
Operations Research, Berkeley.
23

Corsia Carbon Emitions Aviation

Uploaded by

Copyright:

Available Formats

Corsia Carbon Emitions Aviation

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Corsia Carbon Emitions Aviation

Uploaded by

Copyright:

Available Formats

Transportation Research Part D 97 (2021) 102839

Contents lists available at ScienceDirect

Transportation Research Part D

Would CORSIA implementation bring carbon neutral growth in

Available online 23 June 2021

1.1. The US airline industry overview

1.2. Contribution statement and paper structure

2.1. Aviation and the carbon emission

Givoni and Rietveld Global - international - - - ✓ ✓ - - ✓ - ✓ - -

Hofer et al. (2010) United 2004 domestic - - - - - - - - - - ✓ -

Transportation Research Part D 97 (2021) 102839

United ✓ Bureau of This work idenfies the critical

2.2. Factors affecting the GHG emissions in the airline industry

2.3. Carbon Offsetting and Reduction Scheme in International Aviation (CORSIA)

3. Causal relationship between aviation carbon emission and its determinants

3.2. The econometric model

TotalCOit2 = α + β1 FUELit + β2 FUEL/Gt + β3 STAGEit + β4 DEPit + β5 SIZEit + β6 RPM it

s2O ≈ s2B + s2W , (5)

Total COit2 overall 1361995 434279.9 442447.6 2305604 N = 612

FUELit overall 2.19E+08 1.07E+08 2.43E+07 4.93E+08 N = 612

FUEL/Gt overall 2.136454 0.7193239 0.71 4.43 N = 612

STAGEit overall 2369.606 364.7129 1285.568 2986.175 N = 612

DEPit overall 447702.3 143109.5 292139 861482 N = 612

SIZEit overall 208.7892 7.978881 189 220 N = 612

RPMit overall 3.40E+09 2.04E+09 5.68E+08 9.91E+09 N = 612

ADAYSit overall 91707.27 37445.41 42331.26 187632.1 N = 612

LFit overall 0.8026709 0.0529723 0.6234 0.9304 N = 612

EUETS overall 0.4705882 0.4995425 0 1 N = 612

AVGFAREit overall 721.1663 229.1872 329.1468 1530.449 N = 612

3.3. Fixed effect model for panel data regression

3.4. Results of the econometric model

FUELit 0.0032729 0.0041983 − 0.0009254 .

FUELit 0.0033*** 0.0001 34.4200 0.0000 0.0031 0.0035

Notes: Legend: * p<0.05; ** p<0.01; *** p<0.001.

4. The simulation model

4.1. The causal loop diagram: a four point CORSIA framework

4.1.1. Balancing Loop B1

Fig. 3. The causal loop diagram.

4.1.2. Balancing Loop B2

• FlightFrequency⟶ DemandforAirTravel⟶ InternationalPassengers⟶ RevenuePassengerMiles⟶ TotalCO2 Emission⟶

4.1.3. Reinforcing Loop R1

• ProfitMargin⟶ Investment⟶ NewAircraftFleet⟶ FuelEfficiency⟶ FuelConsumption⟶ TotalCO2 Emission⟶

4.1.4. Reinforcing Loop R2

4.2. The stock flow diagram

Fig. 4. The stock flow diagram.

d(Stock)/dt = Inflow(t) − Outflow(t). (7)

4.2.1. Stocks studied under the modeled system

4.2.2. Simulation setup and scenarios

Emission Measure Approach

5.1. Impact of each component on carbon footprints

5.2. Main simulation results

5.2.3. Total emission levels and offsetting requirements

Fig. 5. Impact analysis of different components of CO2 reduction initiatives.

Fig. 6. Simulation results on fuel consumption and emission patterns.

6.1. Dimensional consistency check

6.2. Extreme conditions test

Fig. 7. Simulation results on total emissions and offsetting.

6.3. Statistical validation of simulation results

Correlation Coefficient 0.8348 0.5638

Coefficients 2.1780 3.9041

Notes: Legend: * p<0.05; p<0.01; * p<0.001.