Energy Transition, Climate Change, and COVID-19: Economic Impacts of The Pandemic 1st Edition Fateh Belaïd All Chapter Instant Download

Full download test bank at ebookmeta.
com
Energy Transition, Climate Change, and COVID-19:

Economic Impacts of the Pandemic 1st Edition Fateh
Belaïd
For dowload this book click LINK or Button below
https://ebookmeta.com/product/energy-transition-
climate-change-and-covid-19-economic-impacts-of-
the-pandemic-1st-edition-fateh-belaid/
OR CLICK BUTTON
DOWLOAD EBOOK
Download More ebooks from https://ebookmeta.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Exploring the Consequences of the Covid-19 Pandemic:

Social, Cultural, Economic, and Psychological Insights
and Perspectives 1st Edition Usha Rana
https://ebookmeta.com/product/exploring-the-consequences-of-the-
covid-19-pandemic-social-cultural-economic-and-psychological-
insights-and-perspectives-1st-edition-usha-rana/
Economic Recovery After COVID-19
https://ebookmeta.com/product/economic-recovery-after-covid-19/
Cybersecurity in the COVID 19 Pandemic 1st Edition

Kenneth Okereafor
https://ebookmeta.com/product/cybersecurity-in-the-
covid-19-pandemic-1st-edition-kenneth-okereafor/
The Unequal Pandemic COVID 19 and Health Inequalities

1st Edition Clare Bambra
https://ebookmeta.com/product/the-unequal-pandemic-covid-19-and-
health-inequalities-1st-edition-clare-bambra/
Routledge Handbook of Law and the COVID-19 Pandemic 1st
Edition Joelle Grogan (Editor)
https://ebookmeta.com/product/routledge-handbook-of-law-and-the-
covid-19-pandemic-1st-edition-joelle-grogan-editor/
COVID-19 Pandemic: Lessons from the Frontline Jorge

Hidalgo (Editor)
https://ebookmeta.com/product/covid-19-pandemic-lessons-from-the-
frontline-jorge-hidalgo-editor/
Power, Media and the Covid-19 Pandemic: Framing Public

Discourse 1st Edition Stuart Price
https://ebookmeta.com/product/power-media-and-the-
covid-19-pandemic-framing-public-discourse-1st-edition-stuart-
price/
Stress Tested The COVID 19 Pandemic and Canadian

National Security 1st Edition Leah West
https://ebookmeta.com/product/stress-tested-the-
covid-19-pandemic-and-canadian-national-security-1st-edition-
leah-west/
Pandemics and Travel COVID 19 Impacts in the Tourism

Industry 1st Edition Cláudia Seabra
https://ebookmeta.com/product/pandemics-and-travel-
covid-19-impacts-in-the-tourism-industry-1st-edition-claudia-
seabra/
Fateh Belaïd
Anna Cretì Editors
Energy
Transition,
Climate Change,
and COVID-19
Economic Impacts of the Pandemic
Energy Transition, Climate Change, and COVID-19
Fateh Belaïd • Anna Cretì
Editors
Energy Transition, Climate

Change, and COVID-19
Economic Impacts of the Pandemic
Editors
Fateh Belaïd Anna Cretı̀
Faculty of Management LEDA UMR 760 and Climate Economics
Economics & Sciences, UMR 9221 Chair
LEM-Lille Économie Management Dauphine, PSL University
Lille Catholic University Paris, France
Lille, France
ISBN 978-3-030-79712-6 ISBN 978-3-030-79713-3 (eBook)

https://doi.org/10.1007/978-3-030-79713-3
© Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The impact of COVID-19 pandemic is far beyond health implications. No other

pandemic, including the Spanish Flu of 1918–1920, has had such a powerful impact
on society. Every other aspect of the economy and society is being affected by it,
including economic well-being, contraction in real economic activity, global energy
systems, and the environment. There is a remarkable agreement that the energy
sector, like many others, will evolve in the coming years under the impression
of its consequences. Given the existence of ambitious policy goals aimed at
reducing overall environmental and economic-related crises, the effects of COVID-
19 on energy transition and climate change represent a challenging issue for both
researchers and policymakers as unintended side effects of future conditions.
A range of spectacular policy responses to COVID-19, including travel restric-
tions, business closures, working, and learning from hope, have resulted in a
significant reduction in economic activity and associated fossil fuel usage world-
wide.
Consequently, many countries are claiming significant reductions in greenhouse
gas emissions during 2020, driving them one step closer to the original emissions
targets they committed under the Paris climate change agreement. While the
pandemic may have accelerated progress toward these targets over the past year,
will this trend continue throughout the next decade and beyond?
The unknown scope and duration of the pandemic and its associated economic
shocks have resulted that energy security and clean energy transition becoming
highly unpredictable. As a consequence, the answer to this question will depend,
in part, on the long-term effect of the pandemic on economic activity, global energy
consumption, and the policy instruments which will be implemented to mitigate
climate change.
In this context, this book brings together a number of researchers to discuss the
impact of COVID-19 on energy transition and climate change. It collects a wide
range of high-quality theoretical and empirical studies at the nexus of the COVID-
19 pandemic, energy, resource, and environmental economics. The contributions
collect the most updated data in order to quantify the effects of the pandemics
shocks.
v
vi Preface
Various COVID-19 impacts have been discussed: sustainability issues, potential

for green recovery, air quality, likely long-term effects, the nexus between weather,
pollution, and COVID-19 spread, fuel poverty and energy justice, energy demand
and sustainability in developing countries, including Algeria, Morocco, and Saudi
Arabia. The variety of the topics proposed allows to envisage an interesting range
of policy responses to the different dimensions of the COVID crisis.
Lille, France Fateh Belaïd
Paris, France Anna Cretì

Contents
Assessing the Relationship Between Air Quality, Wealth,

and the First Wave of COVID-19 Diffusion and Mortality . . . . . . . . . . . . . . . . . . 1
Roberto Antonietti, Paolo Falbo, and Fulvio Fontini
COVID-19 Recovery Packages and Industrial Emission
Rebounds: Mind the Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Côme Billard and Anna Creti
Low-Carbon Transition of EU-ETS Firms: Assessing
the Long-Term Effects of Covid-19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Marc Baudry
Modelling Sustainability Transitions Under Covid-19 . . . . . . . . . . . . . . . . . . . . . . . 73
Simona Montagnana and Paolo Zeppini
Weather, Pollution, and Covid-19 Spread: A Time Series
and Wavelet Reassessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Olivier Damette and Stéphane Goutte
The Triple Climatic Dividend of COVID-19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Adel Ben Youssef
COVID-19 and Cognitive Biases: What Lessons Can Be Learned
to Fight Against Global Warming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Michelle Mongo
COVID-19’s Impact on Eliminating Fossil Fuel Subsidies, Hence
Fuel Poverty and Energy Justice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Robin Dickinson
The Impact of the COVID Pandemic: What Can We Expect for
Morocco’s Energy Future? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Maryeme Kettani and María Eugenia Sanin
vii
viii Contents
A Proposition Relationship Between Green Workplace

Environment and Employees Green Behavior on Organizational
and Environmental Impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Lahcene Makhloufi, Fateh Belaïd, and Karima Zidane
The Oil-Price Threshold Effect on External Balances in Saudi
Arabia, Russia, and Canada: Accounting for Geopolitics
and Environmental Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Noha A. Razek, Emilson Silva, and Nyakundi M. Michieka
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Assessing the Relationship Between Air
Quality, Wealth, and the First Wave
of COVID-19 Diffusion and Mortality
Roberto Antonietti, Paolo Falbo, and Fulvio Fontini
1 Introduction
The COVID-19 disease, which spread out at the end of 2019, was declared, on the
11th of March 2020, as a pandemic by the World Health Organization. By April
21, 178 countries had confirmed cases of COVID-19 infection.1 The total count of
reported cases and casualties is still rising at the time of writing this chapter.
The pandemic is having a huge social and economic impact. Social distancing
and lockdowns measures that have been adopted to limit its diffusion have severely
limited industrial, commercial, and transportation activities. On the other hand,
lockdown measures have had positive effects on the environment in general, such
as better air and water quality, less pollution, and a lower anthropic pressure on
several animal species (EEA, 2020). The relationship between the environment and
the COVID-19 pandemic has also attracted attention because it was notable that
the areas being hit the most by the virus were also among the most polluted of the
planet. Wuhan and the province of Hubei, where the outbreak began, Lombardy in
Italy, and the Madrid area in Spain, which have all heavily suffered from the viral
infection, are regions with a very poor air quality.
1 Data obtained from the World Health Organization website: https://www.who.int.
R. Antonietti · F. Fontini ()

“Marco Fanno” Department of Economics and Management, University of Padova, Padova, Italy
e-mail: [email protected]; [email protected]
P. Falbo
Department of Economics and Management, University of Brescia, Brescia, Italy
e-mail: [email protected]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

F. Belaïd, A. Cretì (eds.), Energy Transition, Climate Change, and COVID-19,
https://doi.org/10.1007/978-3-030-79713-3_1
2 R. Antonietti et al.
There is a twofold rationale behind the link identified between air pollution and
the COVID-19 pandemic. First, it has been argued that poor air quality correlates
with a greater diffusion of COVID-19 because atmospheric conditions favoring
the permanence of airborne pollutants, such as particulate matter (PM), would
also facilitate the spread of the virus conveyed by the droplets of human saliva
floating in the air, which seems to be one of the main sources of contagion. PM
could serve as a carrier of COVID-19 virus. Second, there may be a relationship
between air quality and mortality due to COVID-19 infection because chronic
exposure to environmental pollution in general and poor air quality in particular
have a debilitating effect on the body, increasing its exposure to other respiratory
diseases, and reducing the immune system’s response to infections. All these effects
can increase the mortality risk associated with COVID-19.
Research into these aspects is ongoing. By estimating an ecological regression
model on the data of 35 US counties, Wu et al. (2020) provide evidence of the
link between mortality rates and long period exposure to air pollutants (PM 2.5
in particular). Other published studies seem to confirm the above-mentioned links
between air quality and coronavirus diffusion. Wu et al. (2020) estimated an 8%
increase in the COVID-19 death rate associated with a rise of 1 mg/m3 in PM
2.5 levels in some US regions. Ogen (2020) found a positive correlation between
NOx exposure and COVID-19-related mortality in 66 administrative regions in
Italy, Spain, France, and Germany. Setti et al. (2020) found evidence of COVID-
19 on outdoor PM in samples tested in the province of Bergamo (Lombardy, Italy),
which experienced the highest diffusion and mortality rates in Italy (and among the
highest worldwide). Fiasca et al. (2020) have estimated that an increase in PM 2.5
concentration by one unit corresponded to an increase of COVID-19 incidence rates
of 1.56 × 104 people infection due to exposure. Other studies that have found a
significant relationship between PM (2.5. or 10) and COVID-19 cases in Italy are
Fattorini and Regoli (2020) and Bontempi (2020a). However, the evidence gained
to date is not conclusive about the link between air quality and the diffusion of
COVID-19 and the associated mortality (Bontempi, 2020b; Copat et al., 2020), in
particular when taking into account countries’ specificities.
Inspired by the evidence of the uneven diffusion of COVID-19 worldwide, and
supported by recent results, such as in Sarmadi et al. (2020), who find an association
between GDP, meteorological factors, and COVID-19 related variables, we check
whether the macroeconomic structure of countries, as well as more direct factors
like air pollution, plays a role in explaining the first wave of COVID-19 infection
and death rates.
The rationales supporting our conjecture are the following. First, countries’
different levels of wealth can be associated with more or less developed health
care systems, in terms of facilities, personnel, and organization. Wealthier countries
probably had a better chance of taking care of infected people and testing larger
proportions of the population for contagion. This last aspect might also be a
factor introducing a significant measurement bias in the way COVID-19-related
hospitalizations and deaths have been counted.
Assessing the Relationship Between Air Quality, Wealth, and the First Wave. . . 3
At the same time, economic wealth interacts with air pollution levels. We know
that a combination of less efficient production and transport systems, particularly
in less developed countries, and a lower quality of energy consumption coincide
with high environmental externalities (Sovacol, 2012). The cross-country data we
use here confirm as much, showing a negative relationship between real per capita
GDP and air pollution, as measured from the concentrations of small particulate
(PM 2.5).
The role of agriculture needs to be considered as well, not just for its contribution
to GDP, but also because of how it relates to air pollution. At a first glance,
countries based largely on agriculture might be expected to be less exposed to
air pollution, but high-tech and intensive animal breeding is associated with the
extensive use of manure for fertilization, which is in turn associated with large
particulate formation. Our analysis on the COVID-19 pandemic included both
economic and environmental factors, so their interactions were tested too.
The contagion and death percentages observed around the world have been very
different as well as those across the regions inside each country. Much of these
differences is due to idiosyncratic factors, such as the early occurrence/isolation of
a zero patient, the over-/under-evaluation of contagion risks on the side of local
public health authorities, and so on (see, for example, Russo et al. (2020), and
Villaverde and Jones (2020)). Therefore, to assess the relevance of macro socio-
economic factors, whose influence is general and indirect, it is necessary to use
cross section analysis, which smooths out idiosyncratic noise.
To develop our analysis, we merge data on worldwide country-level COVID-19
infections and deaths provided by the European Centre for Disease Prevention and
Control (ECDC) and macro-economic data provided by the World Development
Indicators of the World Bank group. Using a final sample of 142 countries,
we first run a cross-sectional regression using, as the dependent variables, both
COVID-19 infections and deaths, and, as main regressors, air pollution, wealth, and
countries’ total resident population. Then, we cluster countries according to their
“economic similarity” and test the impact of air pollution on COVID-19 infections
and mortality within each group.
This chapter is organized as follows: Section 2 presents the data and the
preliminary analysis; Section 3 presents the cluster analysis and the results of the
related estimates. Section 4 concludes.
2 Data and Preliminary Analysis
To build the dataset, we merge information from two sources. Data on COVID-
19 infections (variable: INFECTIONS) and deaths (variable: DEATHS) are used to
compute the dependent variables. They are drawn from the ECDC, an EU agency
for the protection of European citizens against infectious diseases and pandemics.
The data on the distribution of COVID-19 worldwide are updated on a daily basis
by the ECDC’s Epidemic Intelligence team, based on reports provided by national

health authorities.2
Data for these two variables were collected for 5 days of the first wave of COVID-
19 diffusion: March 24, March 31, April 7, April 14, and April 21, 2020. The
COVID-19 outbreak did not develop everywhere at once, and national authorities
have adopted different strategies and policies to deal with the pandemic. The first
diffusion of COVID-19 has taken a certain amount of time. Countries have reacted
to it with lockdown and other measures that have been implemented differently
across time and countries. All these have affected the measurement of the effects
of stock variables, and that is why we measure the effect of wealth and pollution
on the diffusion and mortality of COVID-19 in different periods, from the start of
the outbreak until the moment when the strictest lockdown measurers started to
be lifted in the European countries have been hit earliest and most severely (Italy
and Spain). At the beginning of our observation period, the relationship might have
been influenced by the different pace at which COVID-19 was spreading around
the globe. By the end of April 2020, lockdown measures were having an effect
on the phenomenon. We nonetheless show stable, significant results across the dates
selected, which means that our findings are robust to the timing of the virus diffusion
and to the heterogeneity of the policies adopted. COVID-19 variables are merged
with the data from the World Bank on:
• PM 2.5: mean annual exposure to PM 2.5 (micrograms per cubic meter).
• GDPPC: real per capita GDP (in 2010 US$ at PPP).
• POPULATION: total resident population.
Table 1 presents the summary statistics of these variables.
Table 2 shows the pairwise correlations of the total number of COVID-19
infections, and the total number of COVID-19-related deaths, with PM 2.5 exposure
and real per capita GDP on 5 different days between March and April 2020.
Table 1 Summary statistics

Variable Mean Std. dev. Min Max
INFECTIONS 24/03 2648.04 10,336.8 1 81,553
INFECTIONS 21/04 17,028.93 72,148.47 6 787,752
DEATHS 24/03 114.98 625.48 0 6077
DEATHS 21/04 1196.82 4956.5 0 42,539
PM 2.5 28.43 20.32 5.861 99.73
GDPPC 15,661.6 20,679.9 370.74 10,9453
POPULATION (mln) 50.554 165.51 0.072 1386.4
2 For more information, see: https://www.ecdc.europa.eu/en/covid-19/data-collection.

Table 2 Correlations of COVID-19 infections and related deaths with PM 2.5 and GDPPC
Infections
24/03 31/03 07/04 14/04 21/04
PM 2.5 −0.059 −0.132 −0.145* −0.143* −0.139*
GDPPC 0.200** 0.275*** 0.276*** 0.269*** 0.261***
Deaths
24/03 31/03 07/04 14/04 21/04
PM 2.5 −0.032 −0.100 −0.148* −0.168** −0.170**
GDPPC 0.107 0.177** 0.243*** 0.274*** 0.280***
***Significant at 1% level; **significant at 5% level; *significant at 10% level
Unlike the literature on air pollution and coronavirus diffusion, we find a negative
correlation between the two, whereas the correlation between coronavirus (both
infections and deaths) and wealth is positive.3 For the number of deaths, we also
note that all correlations become stronger and more significant toward the end of
April.
We test the hypothesis that COVID-19 outcomes (both infections and deaths)
have a significant relation with both PM 2.5 and real per capita GDP (while
controlling for population) by means of a negative binomial regression. The analysis
is replicated for each week from the 24 of March up to the 21 of April; for deaths,
the first week is not considered to take into account the lag between the contagion
and its consequences. Results are reported in Table 3.
We see that the estimated coefficient for PM 2.5 is not statistically significant
when GDPPC is included as a regressor. This indicates that the relationship between
pollution and COVID-19 might be spurious. However, we suspect that the socio-
economic characteristics of each country might play a crucial role in explaining the
link between pollution and COVID-19. To evaluate this, data are integrated with
macro-economic information provided by the World Development Indicators of the
World Bank on:
• IMPORT/GDP: import intensity (i.e., import value as a share of domestic GDP).
• AGRVA/GDP: agriculture value added as a share of GDP.
• MANVA/GDP: manufacturing value added as a share of GDP.4
• CO2 : CO2 emissions (metric tons per capita).
• TEMP: average temperature in March (in ◦ C).
3 We should stress that the correlation between PM 2.5 and COVID-19 infections or deaths is at
country level, or between countries. It may be that, within countries, there is a higher level of
contagion or mortality in regions where air quality is lower.
4 We have omitted the share of services as a proportion of GDP (SERV/GDP) as an explanatory
variable because it is collinear with AGRVA/GDP and MANVA/GDP.

6
Table 3 Infections and deaths: relationship with pollution and wealth

Infections Deaths
NEG BIN (1) (2) (3) (4) (2) (3) (4)
PM 2.5 −0.014 −0.013 −0.012 −0.012 −0.019 −0.020 −0.019
(0.013) (0.009) (0.008) (0.008) (0.018) (0.014) (0.012)
GDPPC 0.00008*** 0.0001*** 0.0001*** 0.00006*** 0.00009*** 0.00008*** 0.00008***
(0.00002) (0.00002) (0.00001) (0.00001) (0.00003) (0.00002) (0.00002)
POPULATION 0.016 0.013 0.012 0.012 0.020 0.018 0.017
(0.016) (0.011) (0.010) (0.008) (0.020) (0.017) (0.014)
N 142 142 142 142 142 142 142
Pseudo R2 0.050 0.043 0.040 0.038 0.053 0.053 0.053
Wald χ 2 23.37*** 31.12*** 34.63*** 36.71*** 30.39*** 36.40*** 41.92***
Alpha 2.753*** 2.455*** 2.394*** 2.350*** 3.862*** 3.699*** 3.585***
AIC 2013.67 2431.55 2553.44 2648.95 1437.10 1567.13 1662.30
(1) = March 24; (2) = April 7; (3) = April 14; (4) = April 21. Robust standard errors in brackets. Each estimate includes a constant term. ***Significant at
1% level; **significant at 5% level; *significant at 10% level
R. Antonietti et al.
Table 4 Summary statistics

Variable Mean Std. dev. Min Max
IMPORT/GDP 0.458 0.250 0.116 1.825
AGRVA/GDP 0.099 0.098 0.0003 0.486
MANVA/GDP 0.128 0.063 0.010 0.374
CO2 per capita 4.954 6.168 0.053 43.86
TEMPERATURE (March, ◦ C) 14.83 11.47 −18.72 30.63
The import intensity measures the degree of (inward) trade openness of the
economy; the agricultural and manufacturing value added considers the different
GDP composition of the economy; The CO2 emissions measure the relative effi-
ciency in using energy as primary energy sources, accounting for both availability of
hydrocarbon primary energy sources and technological development of the energy
sector. Finally, temperature strongly relates with the geographical coordinates of
countries and is suspected to play a crucial role in pandemic diffusions.
Since the World Bank provides information on PM 2.5 exposure up until 2017,
we measure all the explanatory variables in the same year. The CO2 variable has
been included as a measure of the intensity and efficiency with which primary
energy sources are used in a country to generate the aggregate output. Table 4 reports
the summary statistics of these variables.
3 Cluster Analysis
The negative sign of the relationship between the impact of PM 2.5 on infections
and deaths and the possible spurious correlation between COVID-19 and PM 2.5.
deserves further examination. In this section, we check whether the association
between air quality and COVID-19 outcomes changes across different areas of
the world with respect to the economic structure and to climate-related variables.
The 142 countries are grouped using Ward’s method, a well-known hierarchical
approach to grouping observations (see Blashfield, 1980, for example). We identify
seven clusters based on all the variables listed in Table 4. The composition of each
cluster is represented in Fig. 1.
Table 6 reports the list of countries. Table 7 shows the eigenvectors of the
correlation matrix, while the corresponding eigenvalues are shown in Table 8. Table
9 shows the mean values of each item in the clusters.
Cluster 1 is the group that explains the largest amount of the total variance.
It includes many European countries and the USA: these countries share a high-
income level, a large share of services as a proportion of their GDP, a high exposure
to CO2 emissions per capita, a small share of agriculture, and a low temperature in
Fig. 1 Clusters of countries
March. Cluster 2 mainly comprises Eastern Asian and Northern African countries,
which share a low weight of manufacturing as a proportion of domestic GDP and a
moderately high import propensity. Cluster 3 is essentially made up of West Asian
and Sub-Saharan countries, sharing a high openness to imports. Cluster 4 is a mix
of countries sharing high CO2 emissions and a high average temperature in March,
e.g., countries below the Equator. Cluster 5 includes service economies sharing a
low temperature in March. Cluster 6 contains high-income countries specializing
in natural resource extraction. Cluster 7 pools four small open economies (three
islands) with a large share of services and agriculture, and a moderately high level
of CO2 emissions.
We interpret the clusters using a linear discriminant analysis. We thus obtain
linear combinations of the variables (the so-called canonical discriminant axes)
that maximize the separation between the different classes/clusters. These axes are
calculated to respect reciprocal orthogonality, so they can be used to plot individual
data on a Cartesian space, to enable a visual inspection of the bivariate distribution
of the clusters. Figure 2 shows the distribution of the 142 countries, grouped into the
7 clusters (using a different color for each cluster). Four clusters tend to stand out
quite clearly. Clusters 7 and 1, on the right, denote high GDP levels. The countries
they contain exhibit a high share of services as a proportion of GDP, and high levels
of CO2 emissions per capita, a low share of agriculture, and a low temperature
in March. Cluster 7 is also characterized by high import levels. On the left, we
see cluster 3, which is characterized by a high share of agriculture. In the middle,
and lower down, we find cluster 4, with the lowest average import propensity. The
remaining clusters 2, 5, and 6 are not neatly separable on Fig. 2, but Table 9 shows
the average values of each cluster for the variables used in the cluster analysis.
Fig. 2 Distribution of countries and clusters on the first two canonical axes
For each cluster, we adopt seven dummies, which take the value of 1 when
a country belongs to the corresponding cluster. Then, we split our air pollution
variable into seven new variables multiplying PM 2.5 levels by each cluster dummy
(PM2.5i *cluster j). As a final step, we estimate the following equation, one for
COVID-19 infections and one for related deaths (both as on April 21, 2020), using
a negative binomial regression model:

7
YiT = γ0 + γj PM2.5∗i cluster (i)j + β2 POPi + uiT ,
j =1
where T refers to the 21st, 14th, and 7th of April, respectively, and cluster (i)j =1 if
country i belongs to cluster j and 0 otherwise.
Table 5 shows the results. In one case, namely cluster 1 (the wealthiest economies
in the world), higher PM 2.5 concentrations (strongly) correlate with higher rates
of infection and death (at each date). We also find evidence of a (weak) positive
correlation between PM 2.5 and COVID-19 infections in cluster 5. In cluster 3,
the association between PM 2.5 and COVID-19 outcomes is negative and strongly
significant: these low-income countries are mainly in Africa and East Asia. Such
a negative and significant estimated coefficient is nevertheless roughly ten times
smaller than the positive coefficient of PM*cluster1. The same order of magnitude
holds for the marginal effects at the mean: on April 21, a rise of 10 μg in PM 2.5
levels per cubic meter corresponds, on average, to 9850 more infections and 608
more deaths in cluster 1 countries and to 1430 less infections and 64 fewer deaths
in cluster 3 countries.
Table 5 Correlation between pollution, wealth, and COVID-19 infections and deaths, by cluster
Deaths Infections
NEG BIN April 21 April 14 April 7 April 21 April 14 April 7
PM2.5*cluster1 0.304*** 0.313*** 0.333*** 0.211*** 0.223*** 0.234***
(0.053) (0.056) (0.058) (0.045) (0.046) (0.048)
PM2.5*cluster2 −0.013 −0.011 −0.006 −0.020** −0.020* −0.016
(0.013) (0.018) (0.019) (0.010) (0.010) (0.012)
PM2.5*cluster3 −0.032*** −0.031*** −0.025*** −0.031*** −0.030*** −0.032***
(0.009) (0.009) (0.010) (0.007) (0.007) (0.008)
PM2.5*cluster4 0.038 0.043 0.053* 0.028 0.029 0.030
(0.025) (0.026) (0.027) (0.022) (0.023) (0.025)
PM2.5*cluster5 0.033* 0.031 0.031 0.039** 0.036** 0.031*
(0.019) (0.020) (0.020) (0.017) (0.018) (0.018)
PM2.5*cluster6 −0.010 −0.011 −0.010 0.013** 0.009 0.005
(0.008) (0.008) (0.009) (0.005) (0.006) (0.006)
PM2.5*cluster7 0.085 0.064 0.056 0.097* 0.082 0.067
(0.117) (0.107) (0.102) (0.055) (0.064) (0.061)
POPULATION 0.011** 0.010** 0.009* 0.010*** 0.010** 0.010***
(0.005) (0.004) (0.004) (0.004) (0.004) (0.004)
N 142 142 142 142 142 142
Pseudo R2 0.068 0.072 0.078 0.047 0.049 0.051
Wald χ 2 128.1*** 114.4*** 95.96*** 146.1*** 127.8*** 123.1***
Alpha 3.121*** 3.138*** 3.157*** 2.081*** 2.151*** 2.151***
Robust standard errors in brackets. Each estimate includes a constant term. ***Significant at 1%
level; **significant at 5% level; *significant at 10% level. Clusters are identified using the following
variables: GDPPC, IMPORT/GDP, AGRVA/GDP, MANVA/GDP, SERVVA/GDP (services value
added on GDP), CO2 per capita and TEMP
At the same time, there is a limited negative relationship between air pollution
and COVID-19 infections and related deaths for countries in cluster 3, which are
mostly in Sub-Saharan regions (the poorest economies, largely based on agriculture)
for which we cannot advance a plausible explanation. This puzzle might relate to
data quality issues, especially with organizational difficulties and the costs of testing
for the infection on large samples of the population.
4 Conclusions
In this chapter, we have analyzed the relationship between pollution, measured

by concentration of PM 2.5, wealth, and COVID-19 worldwide during the first
wave of the pandemic, taking into account the socio-economic characteristics of
the countries. We have shown that air quality negatively affects both COVID-19
infections and deaths, but this is true only for the richest cluster of countries that are
mostly located in the northern hemisphere. For the other countries, once they are
grouped in different clusters according to their level and composition of GDP, trade
openness, energy efficiency, and climate features, such a relationship does not hold
anymore. This put evidence in favor of the possible linkage between COVID-19
diffusion and pollution through the socio-economic features of the most advanced
countries.
There are several factors to consider regarding the quality of available data on
COVID-19 that can influence our results. The first aspect concerns the homogeneity
of the data collection process. Apart from costs and organizational problems,
different policies have been adopted around the world concerning the use of testing
for the infection and mitigation measures. There has been a generalized scarcity of
test kits, which has influenced how the phenomenon has been measured. Overall,
it is safe to assume that the official COVID counts fall abundantly short of the real
number of infections around the world.
This may be true of the real number of deaths as well. There are non-trivial
problems with certifying a death as being due to COVID-19. It preliminarily
demands doing a test. Many of the elderly people infected with COVID-19 have
been treated outside hospitals and died in nursing homes, adding to the difficulty of
applying the test and establishing the cause of death.5 Besides, a large proportion
of the people dying with the infection are elderly and have underlying medical
conditions, including cardiocirculatory and respiratory problems. In such cases,
definitively establishing the ultimate cause of death is not always easy and can be
costly and time-consuming.
Nevertheless, our analysis gives an account of the impact of air pollution and
economic and environmental variables on the COVID-19 pandemic around the
world. The study of this phenomenon is growing, and we welcome future analyses
that include local and global factors to help explain the relationship between air
pollution and COVID-19 pandemic, as done in this work.
5 In Italy, for instance, the classification protocol states that only people who die after officially
testing positive in hospitals can be classified as COVID-19 victims. Some reports (e.g., Gabanelli
& Ravizza, 2020) show that in several EU countries (e.g., the Netherlands, Belgium, among others),
the mortality rate due to coronavirus remains particularly low, but in the first 4 months of 2020,
these countries have had more than double the mortality rates of the same period in 2019. It is not
clear why different countries count COVID-19-related deaths differently.
Appendix
Table 6 Clusters of countries

Cluster Countries
1 Australia, Austria, Belgium, Canada, Cyprus, Denmark, Spain, Estonia, Finland,
France, Germany, Georgia, Greece, Iceland, Israel, Italy, Japan, Kazakhstan,
Lebanon, Latvia, Montenegro, The Netherlands, Norway, New Zealand, Portugal,
Russian Federation, Sweden, Switzerland, The United Kingdom, The United States
2 Bangladesh, China, Cameroon, Algeria, Gabon, Equatorial Guinea, Guatemala,
Honduras, Indonesia, India, Jordan, Korea, Rep., Morocco, Mexico, Malaysia,
Nicaragua, Oman, Philippines, Paraguay, Senegal, El Salvador, Thailand, Tunisia
3 Afghanistan, Albania, Azerbaijan, Benin, Burkina Faso, Bhutan, Central African
Republic, Cote d’Ivoire, Congo, Dem. Rep., Congo, Rep., Ethiopia, Ghana, Guinea,
Gambia, Guyana, Iraq, Kenya, Cambodia, Liberia, Madagascar, Myanmar,
Mozambique, Mauritania, Niger, Nigeria, Nepal, Pakistan, Rwanda, Chad, Togo,
Timor-Leste, Tanzania, Uganda, Uzbekistan, Vietnam
4 Angola, Argentina, Belize, Bolivia, Brazil, Chile, Colombia, Cabo Verde, Costa Rica,
Cuba, Dominica, Dominican Republic, Ecuador, Egypt, Grenada, Iran, Jamaica, Sri
Lanka, Namibia, Panama, Peru, Uruguay, South Africa, Zambia, Zimbabwe
5 Armenia, Bulgaria, Bosnia and Herzegovina, Belarus, Czech Republic, Croatia,
Hungary, Kyrgyz Republic, Lithuania, Moldova, North Macedonia, Mongolia,
Poland, Romania, Serbia, Slovak Republic, Slovenia, Turkey, Ukraine
6 United Arab Emirates, Bahrain, Brunei Darussalam, Kuwait, Qatar, Saudi Arabia
7 Ireland, Luxembourg, Malta, Singapore
Table 7 Eigenvectors of the Cluster Eigenvalue Difference Proportion Cumulative

correlation matrix
1 3.0519 1.9871 0.4360 0.4360
2 1.0648 0.1769 0.1521 0.5881
3 0.8879 0.1044 0.1268 0.7150
4 0.7835 0.1295 0.1119 0.8269
5 0.6540 0.2849 0.0934 0.9203
6 0.3692 0.1806 0.0527 0.9731
7 0.1886 – 0.0269 1
Table 8 Eigenvalues of the correlation matrix

Cluster → 1 2 3 4 5 6 7
GDPPC 0.480 0.147 −0.014 0.228 −0.045 0.747 −0.370
AGRVA/GDP −0.484 0.199 0.125 0.091 −0.297 0.520 0.587
MANVA/GDP 0.124 −0.835 0.443 −0.120 0.078 0.213 0.161
SERVA/GDP 0.436 0.147 −0.291 −0.373 0.476 0.111 0.572
IMPORT/GDP 0.206 0.463 0.836 −0.076 0.108 −0.158 0.040
CO2 0.390 −0.079 −0.033 0.757 −0.174 −0.280 0.399
TEMP −0.368 −0.004 0.056 0.456 0.797 0.109 −0.078
Table 9 Average values of clusters

Cluster Freq. GDPPC AGRVA/GDP MANVA/GDP SERVA/GDP CO2 IMPORT/GDP TEMP
1 30 39,074 0.03 0.12 0.66 7.9 0.42 2.8
2 23 6217 0.09 0.20 0.53 3.5 0.39 21.1
3 35 1615 0.23 0.09 0.43 0.8 0.43 22.1
4 25 6758 0.07 0.10 0.58 2.6 0.33 22.1
5 19 10,809 0.06 0.16 0.54 5.0 0.61 2.0
6 6 35,377 0.01 0.12 0.52 26.3 0.49 21.8
7 4 66,405 0.01 0.16 0.70 10.1 1.39 10.6
References
Blashfield, R. K. (1980). The growth of cluster analysis: Tryon, Ward, and Johnson. Multivariate
Behavioral Research, 15(4), 439–458. https://doi.org/10.1207/s15327906mbr1504_4
Bontempi, E. (2020a). First data analysis about possible COVID-19 virus airborne diffusion due
to air particulate matter (PM): The case of Lombardy (Italy). Environmental Research., 186,
109639.
Bontempi, E. (2020b). Commercial exchanges instead of air pollution as possible origin of COVID-
19 initial diffusion phase in Italy: More efforts are necessary to address interdisciplinary
research. Environmental Research, 188, 109775.
Copat, C., Cristaldi, A., Fiore, M., Grasso, A., Zuccarello, P., Signorelli, S. S., Conti, G. O., &
Ferrante, M. (2020). The role of air pollution (PM and NO2) in COVID-19 spread and lethality:
A systematic review. Environmental Research, 191, 110129.
EEA. (2020). Air quality and COVID-19, European Environmental Agency. https://
www.eea.europa.eu/themes/air/air-quality-and-covid19
Fattorini, D., & Regoli, F. (2020). Role of the chronic air pollution levels in the Covid-19 outbreak
risk in Italy. Environmental Pollution, 264, 114732.
Fiasca, F., Minelli, M., Maio, D., Minelli, M., Vergallo, I., Necozione, S., & Mattei, A. (2020).
Associations between COVID-19 incidence rates and the exposure to PM2.5 and NO2: A
Nationwide Observational Study in Italy. International Journal of Environmental Research and
Public Health, 17(24), 9318.
Gabanelli, M., & Ravizza, S. (2020). “Morti COVID, tutte le bugie in Europa. Ecco i dati reali”,
Corriere della Sera. Dataroom. https://www.corriere.it/dataroom-milena-gabanelli/morti-
covid-tutte-bugie-dell-europa-ecco-dati-reali/1c28ca00-88b3-11ea-96e3-c7b28bb4a705-
va.shtml (in italian).
Ogen, Y. (2020). Assessing nitrogen dioxide (NO2) levels as a contributing factor to coronavirus
(COVID-19) fatality. Science of the Total Environment, 726, 138605. https://doi.org/10.1016/
j.scitotenv.2020.138605
Russo, L., Anastassopoulou, C., Tsakris, A., Bifulco, G. N., Campana, E. F., Toraldo, G., & Siettos,
C. (2020). Tracing day-zero and forecasting the COVID-19 outbreak in Lombardy, Italy: A
compartmental modelling and numerical optimization approach. PLoS One, 15(10), e0240649.
https://doi.org/10.1371/journal.pone.0240649
Sarmadi, M., Marufi, N., & Moghaddam, V. K. (2020). Association of COVID-19 global
distribution and environmental and demographic factors: An updated three-month study.
Environmental Research, 188, 109748.
Setti, L., Passarini, F., De Gennaro, G., Baribieri, P., Perrone, M. G., Borelli, M., Palmisani, J., Di
Gilio, A., Torboli, V., Pallavicini, A. Ruscio, M., Piscitelli, R., Miani, A. (2020). COVID-19
RNA found on particulate matter of Bergamo in Northern Italy: First preliminary evidence.
MedRXiv. https://doi.org/10.1101/2020.04.15.20065995
Sovacol, B. J. (2012). The political economy of energy poverty: A review of key challenges. Energy
for Sustainable Development, 16, 272–282.
Villaverde, J. F., & Jones, C. (2020). Macroeconomic outcomes and Covid-19: A progress report,
NBER Working Papers 28004. National Bureau of Economic Research.
Wu, X., Nethery, R. C., Sabath, B., Braun, D., & Dominici, F. (2020). Exposure to air pollution and
COVID-19 mortality in the United States: A nationwide cross-sectional study. medRxiv. https:/
/doi.org/10.1101/2020.04.05.20054502
COVID-19 Recovery Packages and
Industrial Emission Rebounds: Mind
the Gap
Côme Billard and Anna Creti
1 Introduction
Since December 2019, the Covid-19 coronavirus has spread quickly from Asia to
Europe and America, causing large-scale loss of life and severe human suffering
(Financial Times, 2020). The pandemic represents the third and greatest economic,
financial and social shock of the twenty-first century—after 9/11 and the global
financial crisis of 2008 (OECD, 2020). From an environmental perspective, this
unexpected episode could mark a turning point in the fight against global warming.
This year, global greenhouse gas (GHG) emissions will fall by around 7%, repre-
senting the annual rate at which our economies should decarbonise to reach carbon
neutrality in 2050.1 Instead, emissions will rebound once mobility restrictions are
lifted and economies recover (Le Quere, 2020), unless governments take actions.
The expected decline in 2020 GHG emissions comes as a consequence of
national policies to prevent the spread of the virus (Helm, 2020).2 Indeed, G20
nations have implemented restrictions (e.g. social distancing, mobility) slowing
down economic systems (Thunström et al., 2020). On the supply side, around 81%
of the global workforce has been hit by either full or partial lockdown measures,
causing unprecedented job losses and furloughs (International Labour Office, 2020).
First draft: June 2020. Authors would like to thank as well the seminar and conference participants
at University of Paris-Dauphine and Climate Economics Chair.
1 This objective ensures a temperature rise below 1.5◦ C degree by 2100 (UNEP, 2019).
2 Suggesting we have not decoupled GDP growth and carbon emissions (Helm, 2020).
C. Billard () · A. Creti

Paris Dauphine University PSL, Paris, France
e-mail: [email protected]; [email protected]
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 15

F. Belaïd, A. Cretì (eds.), Energy Transition, Climate Change, and COVID-19,
https://doi.org/10.1007/978-3-030-79713-3_2
16 C. Billard and A. Creti
On the demand side, consumer spending has fallen as it was no longer possible to
travel, including to shop for discretionary items, go to restaurants or for experience-
based activities (Center for Economic Policy Research, 2020). Overall, the crisis has
impacted economic activity from both demand and supply sides (World Economic
Forum , 2020), calling governments for unprecedented support policies.
Depending on objectives and timing, government support programs are either
part of short-term rescue plans or long-term recovery plans. With respect to
the former, numerous EU governments have already exposed and implemented
fiscal rescue policies (International Monetary Fund, 2020). These emergency
measures come as a short-run safety net to protect business balance sheets, reduce
bankruptcies and address immediate human welfare concerns during lockdown
periods.3 Some rescue policies also cover emissions-intensive companies facing
bankruptcy or significantly reduced revenue. For instance, this has been the case
for airlines companies in France, Australia and the USA.4 To ensure the success
of the Paris Agreement (2015), government support plans (e.g. bailouts) should
be conditional on these industries developing a measurable plan of action to
transition towards a net-zero emissions future.5 Indeed, the COVID-19 crisis will
reduce global GHG emissions in 2020 but the long-term impact of the pandemic
on emissions will be driven by forthcoming investment choices (both public and
private). At the European scale, imminent recovery packages soon to be delivered
will act as stimuli to restore economic growth in the region (Hepburn et al., 2020).
The design of such packages (e.g. sectoral economic incentives, public investments)
will reshape the economy on the long run, acting as a potential game-changer to
reach a post-carbon Union by 2050 (McKinsey & Company, 2020).
A key objective of any recovery package is to stabilise expectations, restore
confidence and to channel surplus desired saving into productive investment
(Hepburn et al., 2020). While most European governments have implemented
rescue packages (e.g. France, Germany, Italy, Spain and the United Kingdom), the
European Commission announced an additional budget amount of e750 billion to
support most affected EU economies, representing more than e1100 billion global
budget for the period 2021–2027 (European Commission, 2020).6 If this amount
is necessary for EU economies to recover (e.g. France, Italy, Spain), an efficient
long-run recovery plan should target sectors able to rapidly create jobs and boost
3 In April 2020, all G20 nations (including most EU member states), had signed such fiscal
measures into law (see International Monetary Fund, 2020).
4 Precisely, France rescue plan for Air France reached e7 billion (Les Echos, 2020); Australian
government announced AU$715mn of unconditional Australian airline relief (through the Coron-
avirus Economic Response Package (Commonwealth of Australia, 2020), and US$32bn of bailouts
for US airlines (see Courtney, 2020 for a review of CARES Act)).
5 Which, for instance, has not been the case for Air France (Le Monde, 2020).
6 Although governments will have flexibility regarding the allocation of such funds, the main
priority is to reach the EU’s objectives of climate neutrality and digital transformation, to offer
social and employment support as well as to reinforce the EU’s role as a global player (European
Parliament, 2020).
COVID-19 Recovery Packages and Industrial Emission Rebounds: Mind the Gap 17
production across other industries in the economy, thus stimulating GDP growth
(Allan et al., 2020). Among other factors,7 those targeted sectors should display high
short-run/long-run economic multipliers, or return for every dollar of expenditure
(Bussing-Burks, 2011; Ramey, 2019). Such metrics accounts not only for the
effects of the spending (government expenses, tax reductions) in the specific sector
(e.g. impact on income, output), but also for the subsequent rounds of spending
generated by the initial expenditures in other parts of the economy.8 Back during
and following the 2008 global financial crisis, expansionary policies, focusing
on investments through the lens of economic coefficients, were more effective at
restarting economic activity than austerity-based policies (European Central Bank,
2015; Hepburn et al., 2020). Twelve years later, the COVID-19 crisis pushes again
policy-makers to decide which key sectors to focus investments on, reflecting
changing technologies and the need to stimulate growth and secure job creations
for the coming years (DG Tresor, 2020).
From a climate policy approach, the sole consideration of economic multipliers
to guide forthcoming economic stimuli does not guarantee a transition towards
a post-carbon society by 2050 (Hammer and Hallegatte, 2020). Indeed, recovery
plans could be either “brown” or “green” depending on their ability to decouple
emissions from economic activity (IFRI, 2020). Sectors exhibiting high economic
multipliers could be those acting as big emitters suggesting a tension between
short-run economic growth and climate targets (see European Commission, 2007).
Then, to decouple GDP growth and emissions, EU governments could be interested
in understanding which industrial sectors are driving GHG emissions9 in the
economy. Following the ongoing crisis, the willingness to design green recovery
plans could lead to not supporting such industries. By doing so, large amounts of
GHG emissions could be avoided once economy recovers, putting the entire EU
industrial system on track with respect to the Paris Agreement objectives. Moreover,
some industries might not be greenhouse gas (GHG) intensive but might decrease
global GHG emissions by limiting provision of inputs to downstream dirty sectors.
When designing sectoral support policies, government will have to be aware of such
intra-sector dynamics to limit emissions. This paper aims at providing new insights
on these issues.
Precisely, we consider the economy as a system of industries interacting with
each other (OECD, 2016) and capture the dynamics of supply/demand between
industrial sectors. Indeed, output of an industry might be used directly as input or
7 Several other factors are relevant to the design of economic recovery packages: contributions to
the productive asset base and national wealth, speed of implementation, affordability, simplicity,
impact on inequality and various political considerations.
8 In detail, economic multiplier measures the impact on activity of each additional currency unit
of spending/tax cut funded by borrowing. A multiplier of 1 means $1 extra spending boosts final
production and income by $1. A multiplier of 3 implies $1 spending boosts final income and output
by $3.
9 Gases that trap heat in the atmosphere (e.g. carbon dioxide, methane), contributing to global
warming. See full description in the Sect. 2.

as output supplied to other sectors (e.g. output of the mining sector are consumed
as inputs by that industry or supplied as inputs to other sectors). The decrease in
production from the mining sector would decrease the sector output and impact the
demand-side sectors. Then, following the chain of intermediate demand, industries
directly connected such as basic metals would in turn reduce their output. Such
a cascading mechanism of output contraction would decrease associated GHG
emissions (i.e. from production).10 While some studies have explored phenomenon
of economic cascades (e.g. information in financial markets (Romano, 2007), the
diffusion of risks in the banking system (Battiston et al., 2017) or the stranding
of dirty assets in a low-carbon economy (Cahen-Fourot et al., 2020)), the topic of
industrial emissions has never been investigated from a systemic perspective.
Overall, we propose a novel analysis of the process through which a contraction
of the gross output of a specific sector would decrease the use of inputs in other
sectors leading to a drop in associated emissions (i.e. forward oriented sectors).
The supply of essential inputs to the rest of the economy is a matter of addressing
primarily forward impact effects rather than backward effects (i.e. change in inputs
affecting upstream sectors). In the context of post-COVID-19 recovery plans,
this approach is particularly relevant as supply chains have been severely hit by
government restrictions, and supply dynamics will be critical to avoid inflation in
the post-COVID opening time-window (BNP Paribas—Economic Analysis, 2020).
While governments will implement economic stimuli to secure high levels of
supply, our paper identifies the sectors that should not benefit from recovery plans11
and quantifies the impacts of such deliberate decision on emissions (i.e. avoided
emissions). Providing such new results would allow policy-makers to account for
these potential “avoided” emissions when designing green economic stimulus. We
use data available for five European countries affected by the COVID-19 pandemic
(i.e. France, Germany, Italy, Poland and Spain) to illustrate our model and achieve
two main objectives.
First, we use Input–Output (IO) concepts to derive national economic matrices
of emission coefficients, including the entire range of the industrial productive
sectors. These coefficients capture the amount of emissions that would be reduced
in a sector due to a unitary decrease in primary inputs12 utilised by another (or
the same) sector, considering both direct and indirect effects. For instance, these
matrices are able to provide GHG emission reductions in the textile sector due to a
drop in the plastics industry, both directly and through its intermediate effects on,
for instance, chemicals. By doing so, we identify industries most likely to trigger
large emission reduction cascades and those most exposed to such a dynamics (i.e.
10 Inthe following, we name “cascading process” such a dynamics of emission contraction.

11 Such sectors drive GHG emissions in the industrial system. Without contributions in terms of
climate strategy, the government willing to achieve climate goals should not target them.
12 We define “primary inputs” as the main factors used in production (labour, capital, land and
others). IO tables report their factor costs (e.g. compensation of employees, consumption of fixed
capital or net operating surplus) (Miller & Blair, 2009).
increase in internal GHG emissions through the channel of another sector). The
novelty of the present analysis is to maintain a systemic perspective of the national
economy, and investigate the transmission channels of GHG emission reductions
across industries (i.e. emission cascades). By providing a quantitative estimation of
such cross-sectoral GHG emission interactions, our paper brings relevant insights to
policy-makers too.
Whatever the economic system, we highlight how mining,13 coke and refined
petroleum products14 and electricity and gas15 are among the sectors with the
largest emission coefficients. A fall in their activity (i.e. gross output) creates the
largest reduction amount of emissions in the system. Leaving all else equal, green
recovery packages should ensure their activity to not expand, and even further
to contract.16 On the opposite, coke and refined petroleum, basic metals17 and
electricity and gas industries are the most exposed to such dynamics of emission
contraction. All these activities have various impacts in terms of GHG emission
reductions (across economies), suggesting different national strategies regarding
implementation of recovery plans.
Second, we focus our study on the mining industry in order to investigate the
most relevant channels of sectoral cascades of GHG emissions that could be avoided
in the future. By doing so, we are able to evaluate the role of energy intensive
sectors such as coke and refined petroleum products, basic metals and electricity and
gas in the cascading process. While countries exhibit different cascading dynamics
depending on the peculiarities of their industrial structure, certain regular patterns
emerge. On one hand basic metals, the manufacture of coke and refined petroleum
products and electricity and gas are the activities the most directly exposed to a
drop in GHG emissions through a contraction of production in the mining sector.
Such results suggest strong economic connections between the mining industry and
these sectors where the former supplies the latter.18 On the other hand, irrespective
to their rankings in the process, chemicals and pharmaceutical products19 as well
13 The sector encompasses coal and lignite, crude petroleum and natural gas, metal ores, other
mining and quarrying products and mining support services.
14 Includes coke oven products and refined petroleum products.
15 The sector mainly covers electricity, transmission and distribution services, manufactured gas,
distribution services of gaseous fuels through mains, steam and air conditioning supply services,
natural water, water treatment and supply services.
16 If one assumes no shift towards cleaner production in those industries.
17 The sector covers basic iron and steel and ferro-alloys, tubes, pipes, hollow profiles and related
fittings, of steel, other products of the first processing of steel, basic precious and other non-ferrous
metals.
18 In some countries such as Germany and Poland, this finding is particularly relevant.
19 Sector covers basic chemicals, fertilisers and nitrogen compounds, plastics and synthetic rubber
in primary forms, pesticides and other agrochemical products, paints, varnishes and similar
coatings, printing ink and mastics, soap and detergents and other chemical products.
as manufacture of other non-metallic mineral products20 are highly present in the

third layer of the GHG cascading process.21 The latter emphasises the existence of a
significant connection between those activities and aforementioned energy intensive
sectors (e.g. chemicals affected by upstream coke and refined petroleum products
activity). In addition, agriculture22 and construction23 are often impacted by mining
decreasing activity through the channel of basic metals. From a policy perspective,
these outcomes suggest that moving away from mining would have impacts not
only on emissions (Fugiel et al., 2017), but would also generate economic effects
on other sectors (i.e. from mining to downstream industries such as construction
through the channel of basic metals). On this issue, our results complement the
flourishing literature on assets at risk due to a low-carbon transition (Creti & de
Perthuis, 2019). Such findings reinforce the importance of a well-designed recovery
plan to limit GHG emission rebound effects and to stimulate sector-based clean
solutions (e.g. green inputs).
The remainder of the article is organised as follows. Section 2 introduces the
method to compute the matrices of sectoral emission coefficients. Section 3 presents
the results of the analysis for five European countries, discussing the sectors most
likely to create large amounts of emission reductions and the ones most exposed
to such dynamics. Section 4 focuses on understanding the systemic propagation of
shocks starting from mining in order to identify relevant channels of GHG emission
decline. Finally, Sect. 5 discusses implications of our results for designing effective
green recovery packages to avoid a resurgent increase in industrial GHG emissions
and exposes elements of conclusion.
20 The activities include glass and glass products, refractory products, clay building materials, other
porcelain and ceramic products, cement, lime and plaster, articles of concrete, cement and plaster,
cut, shaped and finished stone.
21 For non-metallic mineral products, the sector is present in the second or third layer, depending
on the examined country.

22 The sector includes non-perennial crops, perennial crop, planting material: live plants, bulbs,
tubers and roots, cuttings and slips, mushroom spawn, live animals and animal products,
agricultural and animal husbandry services (except veterinary services), hunting and trapping and
related services, forest trees and nursery services, wood in the rough, wild growing non-wood
product, support services to forestry, fish and other fishing products; aquaculture products, support
services to fishing.
23 Represents buildings and building construction works, roads and railways, construction works
for roads and railways, constructions and construction works for utility projects; constructions and
construction works for other civil engineering projects, demolition and site preparation works;
electrical, plumbing and other construction installation works, building completion and finishing
works, other specialised construction works.
2 Methodology and Data
2.1 The Emission Reduction Multiplier Matrix
Following Cahen-Fourot et al. (2020), we start with the national inter-industry

matrix Z, a square matrix exhibiting amounts of sectoral intermediate consumption.
In broad terms, such a matrix is called “Input–Output matrix” and captures
exchanges of goods and services among industrial sectors in monetary units.24
In input–output tables (IOTs), the Z matrix usually comes with an additional set
of column vectors displaying final consumption (i.e. demand (f)) and row vectors
representing value added items (v) (i.e. compensation of employees, fixed capital
consumption, gross operating surplus). Sectors appear both as producers of goods
and services (rows) and as consumers of intermediate inputs (columns). More
specifically, IOTs are commonly defined as monetary industry balances, where total
supply xT = iT Z + v equals total use x = Zi + f of products and services per
sector.25 Therefrom, the total amounts of all transactions over a row (industry output
allocated to each category of user (i.e. intermediate and final consumption)) equals
the sum over the corresponding column (total industry input flowing from upstream
sectors—other industries and value added items). The IOT also reports imported
goods and services, which are again used either as intermediate inputs or as final
demand. Figure 1 below shows a stylised version of an IOT.
Fig. 1 A stylised input–output table (Cahen-Fourot et al., 2020)
24 See Miller and Blair, 2009; Cahen-Fourot et al., 2020.

25 Note that i is a column vector of the same dimension of Z.
In economics, IOTs have been mainly used to evaluate direct and indirect effects
of changes in final demand based on the Leontief inverse matrix (Leontief, 1951;
Metzler, 1951; Chen, 1973; Velàzquez, 2006). Although the demand will be critical
in defining the forthcoming dynamics of GHG emissions under recovery plans, the
novelty of our analysis is to adopt a supply-side perspective. Namely, we capture
those sectors providing lower amounts of inputs supplied to other sectors as a
result of a one-unit decrease in their gross value added or, generally speaking, gross
domestic product (i.e. forward oriented sectors); this allows us to capture associated
decrease in emissions (described hereafter). The supply of essential inputs to the rest
of the economy is a matter of addressing primarily forward impact effects rather than
backward effects (i.e. change in inputs affecting upstream sectors). In the context of
post-COVID-19 recovery plans, our methodology is relevant as the supply has been
particularly hit by governments restrictions (e.g. mobility) and supply dynamics
will be critical to meet forthcoming demand and avoid inflation (BNP Paribas—
Economic Analysis, 2020). Governments, while implementing economic policies
to support supply-side sectors, will have to be careful on the potential impact on
GHG emissions.
A relevant approach for our study is the Ghosh (1958) supply-driven
model.26 The output of the Ghosh model is a matrix B = x−1 Z of allocation
coefficients of the supply of a sector (i.e. output) to all other sectors. In the matrix
B, each element bij quantifies the share of sector i’s output consumed by sector j.
Then, the Ghosh matrix G is defined as:
G = (I − B)−1
We then transpose G to be able to read the effects of changes in sectoral primary

inputs over the columns (similarly to the Leontief system) of GT , where T denotes
the matrix transposition. Each entry gi,j of GT shows the change in output x in
sector i that would result from a unitary change of primary inputs used in sector j. In
general terms, a drop (or increase) of one monetary unit in primary inputs supporting
production in sector i will generate a drop (increase) in the output of sector j by an
amount equivalent to gi,j .27 In IO analysis, primary inputs cover items appearing
on the rows below the inter-industry matrix (e.g. compensation of employees). As
exposed in Cahen-Fourot et al. (2020), primary inputs represent the societal effort
to produce the output of a sector, captured by factor payments.
We innovate by combining the obtained Ghosh matrix with sectoral data of
GHG emissions e.28 To do so, we define Ei = ei /Mid as the emission intensity
of sector i, where M d represents the domestic output of the sector. By multiplying
the diagonalised form of the vector of emission intensities by the Ghosh matrix, we
find the matrix S of emission reduction coefficients:
26 Augustinovics, 1970; Beyers, 1976.

27 Note that gi,j includes both direct and indirect effects.
28 Cf. next part for full description of data.
S = ÊGT
Each element sij of matrix S represents the change in emissions in sector i generated
by a unitary change of primary inputs used by sector j. For our purpose, the elements
of S capture the amount of emissions of a sector i that could be reduced because of
a unitary decrease in primary inputs used in the production of goods and services
of another sector j (e.g. hard coal, iron ores). The column sum of matrix S gives a
measure of the total amount of reduced emission resulting from a unitary reduction
of primary inputs in a sector j. We define this as the total emission coefficient of a
sector:
sjT OT = iT S
where n is the dimension of matrix S. In our case, we assume the values of siT OT
to be largely driven by i emission intensity and therefore, by potential amounts of
internal emission reduction. On the opposite, to estimate external emission reduction
coefficient (i.e. the impacts of a sector reduction of primary inputs on emissions of
all other sectors), we proceed as follows:
diag
sjEXT = sjT OT − sj
where s diag refers to the j-th element of the diagonal of S. In the end, we define
the sum of the rows of S as the exposure of a sector to emission reductions (i.e. the
reduction of emissions following a unitary loss in primary inputs used in all other
sectors):
siEXP = Si
Overall, this methodological approach allows us to investigate both internal

and external emission reductions generated by sectors. A sector might have large
emission reduction coefficients mainly driven by internal reductions—suggesting a
poor economic connection with other sectors (supply). We capture this feature in
Sect. 3 by constructing the channels of reduction cascades across economies.
2.2 Datasets: Input–Output Tables and Emissions
We apply the methodology described above to five European economies, heteroge-

neously affected by the pandemic crisis: France, Germany, Italy, Poland as well as
Spain (OECD, Economic Outlook, 2020; The Guardian, 2020). The main source of
Table 1 Breakdown of examined NACE sectors

Sector Code Sector description
A 1 Agriculture, forestry and fishing
B 2 Mining and quarrying activities
C10–12 5 Food products, beverages and tobacco
C13–15 6 Textiles, wearing apparel, leather and related products
C16 7 Wood and of products of wood and cork (except furniture)
C17–18 8 Paper products and printing
C19 9 Coke and refined petroleum products
C20–21 10 Chemicals and pharmaceutical products
C22 11 Rubber and plastics products
C23 12 Other non-metallic mineral products
C24 13 Manufacture of basic metals
C25 14 Fabricated metal products, except machinery and equipment
C26 15 Computer, electronic and optical products
C27 16 Electrical equipment
C28 17 Machinery and equipment n.e.c.
C29 18 Motor vehicles, trailers and semi-trailers
C30 19 Other transport equipment
C31–33 20 Other manufacturing, repair and installation of machinery and equipment
D–E 21 Electricity, gas, water supply, sewerage, waste and remediation services
F 22 Construction
IO tables data we employ is extracted from the OECD for the year 2015.29 More
precisely, we use symmetric input–output tables at basic price by industry.30 Table 4
in the Appendix lists NACE level 1 categories,31 while Table 1 below offers the
detailed disaggregation of industries we investigate in this paper.32 In the following,
we deliberately exclude business services (Table 4, from G to S) as it represents a
small share of emissions (OECD, Air Emission Accounts).33 However, by providing
inputs to other sectors, such activities can still play a significant role by driving down
emissions. Our model allows us to capture such dynamics through the channel of
external coefficients. With respect to emissions, we constructed our dataset from
the OECD—Air emission accounts on total GHG emissions per sector (CO2 eq.)
29 OECD Statistics: https://stats.oecd.org/Index.aspx?DataSetCode=IOTSI4_2018.

30 Total economy, product by product in million $.
31 The statistical classification of economic activities in the European Community, abbreviated as
NACE, is the classification of economic activities in the European Union (EU).

32 For further descriptions, refer to the Appendix.
33 Exposed by Cahen-Fourot et al. (2020): The decarbonisation process might not be particularly
detrimental for services activities (low dirty capital levels, low demand for fossil fuel).
for the year 2015.34 Overall, Germany emitted more than 629 Mt (CO2 eq.),
followed by Poland (297 Mt), Italy (258 Mt), France (247 Mt) and Spain (221
Mt).35 Although countries exhibit different patterns in terms of sectoral emissions,
our dataset suggests that agriculture, electricity and gas, chemicals, fabricated
metal products and basic metals were the largest emission intensive activities
in 2015.
By combining IOTs and data of emissions, we are able to offer results for the
industrial and power sectors (i.e. NACE from A to F) in France, Germany, Italy,
Poland and Spain. In 2015, these countries represented more than 60% of the
European Union gross domestic product (Statista, 2020).
Amount of GHG emissions (Mt CO2 eq.)
A
D-E
300 C23
200
100
0
e
d
y
in
c
an
al
an
a
an
It
Sp
m
l
Po
Fr
er
G
Distribution of total GHG emissions (CO2 eq.) across Agriculture (A), Electricity and Gas (D–E)
and Other Non-metallic mineral products (C23)
34 OECD Statistics—“Data refer to total emissions of CO2 (CO2 emissions from energy use and
industrial processes, e.g. cement production), CH4 (methane emissions from solid waste, livestock,
mining of hard coal and lignite, rice paddies, agriculture and leaks from natural gas pipelines),
N2O (nitrous oxide), HFCs (hydrofluorocarbons), PFCs (perfluorocarbons), (SF6 + NF3) (sulphur
hexafluoride and nitrogen trifluoride), SOx (sulphur oxides), NOx (nitrogen oxides), CO (carbon
monoxide), NMVOC (non-methane volatile organic compounds), PM2.5 (particulates less than
2.5 μm), PM10 (particulates less than 10 μm) and NH3 (ammonia)”.
35 These data cover the scope of our analysis (e.g. services are not included, imported emissions
neither).
3 Cascading Greenhouse Gas Emissions
3.1 Emission Coefficients
In this section, we analyse results reported in Table 2 below, namely total emission
coefficients (1), external coefficients (2) and exposure to emissions (3). We focus
on the top five sectors for each country. Given the distribution of emissions in the
economy and leaving all else equal, the first two sets of coefficients show the sectors
that are likely to generate the largest amounts of GHG emission reductions (Mt CO2
eq.) in the economic system following a unit decrease in their primary inputs. For
the purpose of our paper, these sectors are the ones that should not be supported
by forthcoming recovery plans if governments are willing to decouple growth and
emissions. On the opposite, recovery plans could create economic incentives to
engage such sectors in cleaner production processes (cf. Sect. 5). The third set
of results displays instead the sectors that are likely to be most exposed to such
dynamics of decreasing emissions from a unitary drop36 distributed equally across
all industries.
Table 2 Emission coefficients

France Germany Italy Poland Spain
Total emission coefficients (1)
B (0.0292) B (0.0190) B (0.0373) B (0.0121) B (0.0286)
C19 (0.0058) D–E (0.0055) C19 (0.0081) D–E (0.0118) C19 (0.0061)
C24 (0.0035) A (0.0051) D–E (0.0044) C19 (0.0073) C23 (0.0057)
A (0.0030) C19 (0.0044) C23 (0.0032) C24 (0.0062) D–E (0.0040)
C23 (0.0028) C24 (0.0027) C24 (0.0028) A (0.0054) C24 (0.0027)
External emission coefficients (2)
B (0.0287) B (0.0178) B (0.0359) B (0.0088) B (0.0280)
D–E (0.0004) C19 (0.0008) M (0.0007) C20–21 (0.0019) C20–21 (0.0008)
C19 (0.0004) G (0.0005) K (0.0007) C19 (0.0016) D–E (0.0008)
C22 (0.0004) M (0.0004) C24 (0.0006) C24 (0.0015) C22 (0.0007)
C25 (0.0004) C25 (0.0004) C19 (0.0006) C28 (0.0013) K (0.0006)
Exposure to emission coefficients (3)
C19 (0.0225) D–E (0.0114) C19 (0.0268) D–E (0.0142) C19 (0.0205)
D–E (0.0042) C19 (0.0083) D–E (0.0141) C19 (0.0058) D–E (0.0104)
C24 (0.0029) C24 (0.0036) C24 (0.0026) A (0.0053) C24 (0.0039)
A (0.0026) A (0.0017) C23 (0.0016) C24 (0.0029) C23 (0.0031)
C23 (0.0012) C23 (0.0017) C20-21 (0.0009) C23 (0.0021) A (0.0015)
36 In primary inputs.
Regarding total coefficients, sector of mining (B) is by far the most prevalent,
appearing as the top sector in every country of our scope.37 Studying the S matrix,
one can notice that emissions of sectors often significantly affected by the drop
in primary inputs originating in the mining industry include those from coke and
refined petroleum products (C19), electricity and gas industry (D–E), other non-
metallic mineral products (C23) and basic metals (C24). These results emphasise
the critical presence of mining inputs in their production process (e.g. iron ores,
coal). For certain countries (e.g. Germany, Italy, Poland), the presence of D–E sector
is likely to be mainly driven by the proportion of energy producing inputs in the
energy mix (e.g. coal/gas, see EU data, Energy statistical datasheets for the year
2015).
In addition to B activities, industries included in category C (manufacturing)
such as C19 (coke and refined petroleum products), C23 (other non-metallic mineral
products), C24 (basic metals) and D–E (electricity and gas) exhibit large coefficients
of emission reductions across economic systems.38 The latter appears to be strongly
intertwined with the level of emission intensity of the sectors, thus highlighting
a significant potential for an internal emission decline. Moreover, for the specific
case of other non-metallic mineral products (C23), examined EU countries were
the largest producers in the EU in 2015 (European Commission, 2017). The strong
potential for internal emission contraction in these industries is confirmed by the
following analysis on external emission coefficients. Finally, agriculture (A) is
among the top sectors of total emission coefficients in France, Germany and Poland.
This outcome emphasises the key role of agricultural practices in climate mitigation
strategies (IPCC, 2014).
External emission coefficients, which abstract from internal emissions of a sector
and thus offer an accurate representation of the effect of a sector’s activity on
GHG emission decline in the rest of the economy, exhibit a different pattern. The
relevance of mining (B) as an import-intensive activity is still highly significant
(i.e. coefficients), confirming a strong economic connection (i.e. provider of inputs)
between this sector and other high polluting sectors (e.g. coke and refined petroleum
products (C19) and electricity and gas (D–E)).39 With respect to other GHG
intensive sectors, coefficients are drastically reduced. Sectors C20–21 (chemicals
37 Remember that total coefficients are column sums of the S matrix, thus representing the
cumulative impact of a drop in a sector’s primary inputs on GHG emissions of other sectors.
As to interpret the coefficient of mining: a one-unit decrease (in monetary unit = million $) in
mining primary inputs leads to a drop in GHG of 0.029 Mt (CO2 eq.) across all other sectors in
the economy. Looking at the S matrix and the impact of mining on coke and refined petroleum
products (C19) we have: a one-unit decrease (in monetary unit = million $) in mining primary
inputs leads to a drop in GHG from the coke and refined petroleum industry of 0.020 Mt (CO2
eq.).
38 Note that for France, D–E is not among top sectors. We expect the latter to be due to the large
share of nuclear power generation in the country.

39 Note that mining external coefficients are significantly high, embodying the ability of the sector
to generate emissions in other GHG intensive sectors. Moreover, mining products are mainly
imported from outside of the EU, thus explaining low amounts of emissions for the sector (although
and pharmaceutical products) become particularly relevant in Poland and Spain

while C19 (coke and refined petroleum products) and C24 (basic metals) are the
most recurrent manufacturing sectors in our sample (except in Spain). All these
sectors appear high in the ranking of external emission coefficients because they
provide significant amount of inputs to other productive sectors (thus driving up
emissions).40 For instance, both sectors B (mining) and C20–21 (chemicals) provide
substantial intermediate goods to coke and refined petroleum products (C19) while
basic metals (C24) supply fabricated metal products (C25) as well as machinery and
equipment (C28). Table 3 below offers a closer look at industries exhibiting largest
coefficients, reporting the top 5 sectoral values for the external emission coefficients
originating in mining (B).41
As mentioned, C19 (coke and refined petroleum products), C24 (basic metals)
and D–E (electricity and gas) all appear as the sectors most exposed to a decrease
in GHG emissions through the channel of mining. Again, this matches previous
observations that top industries in external emission coefficients provide substantial
inputs to GHG intensive sectors.42 Moving back to Table 2, several other manufac-
turing sectors appear among the top 5. For instance, this is the case for activities C22
(rubber and plastics) and C25 (fabricated metal products) in France, Germany and
Spain, respectively. Note that financial and insurance activities (K) are also present
in Italy and Spain.
Finally, looking at the values of total sectoral exposure to emissions, we can
identify four main sectors, repeatedly appearing among the sectors with the highest
row sums in S: C19 (coke and refined petroleum products); C23 (other non-metallic
mineral products); C24 (basic metals) and D–E (electricity and gas). These sectors,
in addition to having high emission intensities, are affected by multiple relevant
inward economic links. To investigate these features, we consider the S matrix as an
Table 3 Sectoral emission coefficients for top sectors (excluded)

France Germany Italy Poland Spain
B (0.0292) B (0.0190) B (0.0373) B (0.0121) B (0.0286)
C19 (0.02094) C19 (0.0072) C19 (0.0236) D–E (0.0038) C19 (0.0171)
D–E (0.0032) D–E (0.0070) D–E (0.0095) C19 (0.0027) D–E (0.0061)
C24 (0.0018) C24 (0.0022) C24 (0.0012) C24 (0.0010) C24 (0.0023)
A (0.0009) C23 (0.0004) C23 (0.0005) A (0.0004) C23 (0.0012)
C23 (0.0007) A (0.0003) C20–21 (0.0002) C23 (0.0004) C20–21 (0.0005)
in Poland, the sector displays a high amount of emissions as the country is the biggest EU hard-coal
producer Reuters, 2020).
40 The low level of external emissions coefficients of energy intensive sectors is due to the fact
that downstream sectors are not huge polluting industries (e.g. machinery and equipment (C27),
construction (F)).
41 We exclude respective country top sector itself, to abstract from internal emissions.
42 Although for electricity and gas, this argument depends on considered energy used (gas, coal).
adjacency matrix for a directed network (Godsil & Royle, 2013; Halleck-Vega et al.,
2018; Cahen-Fourot et al., 2020), interpreting productive sectors as the vertices of
the network and the si,j elements of S as the weight of the edges going from vertex j
to vertex i. Then, it is possible to represent the network as a circular layout. Figures 2
and 3 below show the outcome of this procedure for Germany and Spain, as they
exhibit different patterns in terms of sectors’ exposure (e.g. strength of coefficients
and rankings).
When studying networks’ characteristics, if one considers the potential strength
(weight) of forward emission links,43 mining activities (B) as well as coke and
refined petroleum products (C19) and fabricated metal products (C25) are the
leading sectors in Germany while mining (B) together with chemicals (C20–21)
and electricity and gas (D–E) display the largest impacts in Spain. Interestingly,
in Germany the most important GHG emission links start from B to coke and
refined petroleum products (C19) and electricity and gas (D–E). The latter could
be partly explained by the coal and gas dominating roles in the German power
generation system (International Energy Agency, Key energy statistics, 2018). For
Spain, top emission connections follow the same pattern: They start from mining
Fig. 2 Network of emissions across sectors in Germany. The size of the node is proportional to
the number of weighted incoming links
43 Thus suggesting a strong effect of reducing gross domestic output in these sectors on other
sectors’ GHG emissions.
Fig. 3 Network of emissions across sectors in Spain. The size of the node is proportional to the
number of weighted incoming links
(B) towards coke and refined petroleum products (C19), electricity and gas (D–E)
and basic metals (C24). Comparing both countries, one can note on Figs. 2 and 3 the
significant weight of emission links from mining (B) to electricity and gas (D–E),
coke and refined petroleum products (C19) and basic metals (C24) in Germany. The
latter provides a clear view of the strong role of mining (B) in the industrial German
ecosystem.
Overall, our results emphasise different aspects: First, mining (B) is the sector
most able to drive down external emissions, exhibiting strong links/connections to
high polluting activities.44 The sector deeply affects coke and refined petroleum
products (C19) as well as basic metals (C24) and electricity and gas (D–E)
industries. For this reason, we observe these industries to be ranked top in
exposure category. Moreover, the ranking of coke and refined petroleum products
(C19), chemicals (C20–21) as well as basic metals (C24) among the top external
emission activities suggests a strong connection of these sectors to others across the
economy—thus acting as facilitators in the shock propagation process originating
from mining (B). The strength of such edges informs policy-makers not only on the
dependence of sectors to others, but also on the ability of key industries to reduce
emissions elsewhere in the economy.
44 Although mining is not emitting large amounts of emissions, cf. Table 5 in the Appendix.
In the next section, we investigate this feature. We map the cascade of GHG
emission contractions from top external coefficients activities. By doing so, we
are able to capture the key sectors acting as drivers of emission reductions in the
industrial system. The latter brings us a clearer perspective on the existence of
common or various patterns of cascades of emission contractions across countries.
4 Channels of Emission Cascades
After having shown the emission potential and associated exposure for the entire
range of productive sectors, we shift our attention to top external multiplier activities
(B activities). Our objective is to better investigate the propagation channels of
decreasing GHG emissions due to a contraction (gross output) originating from
the fossil fuel industry (e.g. coal and gas). Precisely, we trace out the propaga-
tion process throughout the industrial system to capture relevant patterns across
economies.
We start by identifying the most relevant emission links resulting from a unitary
drop of primary inputs supporting the production of mining (i.e. the largest values
appearing on the B column of matrix S). We retain only the top q percentile of the
values and position the affected sectors on the first layer of our cascade network. We
repeat the procedure for the sectors in the first layer, identifying the sectors within
the top q percentile of emissions originating in the layer. The weight of the resulting
network edges is re-weighted to take into account that the fall in primary inputs in
these sectors will be lower than one and a function of the strength of the upper edges.
In other words, the emission reduction links tend to be stronger the closer they are
to the shock origin, and get gradually weaker as they cascade downwards. We then
repeat the same procedure for each layer, excluding the sectors that had already
appeared in upper layers, until no new sectors appear. The results of this procedure
are shown for each country belonging to our sample following a hierarchical layout
(cf. Figs. 4, 5, and 6 below, for q = 0.2). The numerical weight of the top 10 edges
is shown for reference.45
As expected, the sectors in the first layer of the network overlap with the
ones reported in Table 2. The strongest emission link is the one flowing from
mining (B) to coke and refined petroleum products (C19) for France, Germany,
Italy and Spain. Interestingly, the reduction link from mining (B) to electricity
and gas (D–E) has a larger weight in Poland while reaching an identical level
compared to coke and refined petroleum products (C19) in Germany. The latter
confirms previous observations on the carbon intensity of power systems in those
economies. Manufacturing activities, especially basic metals products (C24), other
non-metallic mineral products (C23) as well as electricity and gas (D–E), also
frequently appear among the sectors most strongly affected by the immediate
45 Althoughmost of them exhibit a weight of 0, the impact on downstream sectors remains higher
compared to other industries.
Fig. 4 Hierarchical networks of emission cascades across economic sectors in France (left) and
Germany (right)
Fig. 5 Hierarchical networks of emission cascades across economic sectors in Italy (left) and
Poland (right)
contraction caused by mining (B). From the electricity and gas sector (D–E), the
emission cascade often continues, further affecting chemicals (C20–21) or other
non-metallic mineral products (C23) (cf. Germany, Italy, Poland, Spain). Given the
strength of the original emission connection from mining (B) to electricity and gas
(D–E), these links are often the most relevant after the ones affecting sectors in
the first layer, and are justified by both the high emission intensity of the sectors
and their large consumption of energy products (e.g. electricity, gas). From coke
and refined petroleum products sector (C19), the most common cascades proceed
through the chemical sector (C20–21)46 while from basic metals sector (C24), GHG
emission cascade propagates through the construction sector (F) and agriculture (A).
46 Note that it is relevant in every country of our sample.

Another random document with
no related content on Scribd:
TO CLEAR VEGETABLES FROM INSECTS.
Lay them for half an hour or more into a pan of strong brine, with
the stalk ends uppermost; this will destroy the small snails and other
insects which cluster in the leaves, and they will fall out and sink to
the bottom. A pound and a half of salt to the gallon of water will
answer for this purpose, and if strained daily it will last for some time.
TO BOIL VEGETABLES GREEN.
After they have been properly prepared and washed, throw them
into plenty of boiling water which has been salted and well skimmed;
and keep them uncovered and boiling fast until they are done, taking
every precaution against their being smoked. Should the water be
very hard, a small half-teaspoonful of carbonate of soda, may be
added with the salt, for every two quarts, and will greatly improve the
colour of the vegetables; but if used in undue proportion it will injure
them; green peas especially will be quickly reduced to a mash if
boiled with too large a quantity.
Water, 1 gallon; salt, 2 oz.; soda, 1/4 oz.; or carbonate of soda, 1
teaspoonful.
POTATOES.
(Remarks on their properties and importance.)

There is no vegetable commonly cultivated in this country, we
venture to assert, which is comparable in value to the potato when it
is of a good sort, has been grown in a suitable soil, and is properly
cooked and served. It must be very nutritious, or it would not sustain
the strength of thousands of people whose almost sole food it
constitutes, and who, when they can procure a sufficient supply of it
to satisfy fully the demands of hunger, are capable of accomplishing
the heaviest daily labour. It may not be wise to depend for
subsistence on a root of which the crop unhappily is so frequently in
these days destroyed or greatly injured by disease, and for which it
is so difficult to find a substitute that is equally cheap, wholesome
and satisfying; but we can easily comprehend the predilection of an
entire people for a tuber which combines, like the potato, the solidity
almost of bread, with the healthful properties[103] of various other
fresh vegetables, without their acidity; and which can also be cooked
and served in so many different forms. The wretched manner in
which it is dressed in many English houses renders it comparatively
valueless, and accounts in a measure for the prodigality with which it
is thrown away when cold, even in seasons when its price is highest.
[104]
103. The late Dr. Pereira has stated in his excellent work on diet, page 370, that
Dr. Baly, who has published some interesting observations on the anti-
scorbutic quality of the potato, says, “As ordinarily cooked, it is an admirable
preservative against the scurvy,” for which it appears to be also a cure, see
the same work.
104. We cannot refrain from a few words of remark here on the daily waste of
wholesome food in this country which constitutes one of the most serious
domestic abuses that exist amongst us; and one which it is most painful to
witness while we see at the same time the half-starvation of large masses of
our people. It is an evil which the steady and resolute opposition of the
educated classes would soon greatly check; and which ought not vainly to
appeal to their good sense and good feeling, augmenting, as it must, the
privations of the scantily-fed poor; for the “waste” of one part of the
community cannot fail to increase the “want” of the remainder.
TO BOIL POTATOES.
(As in Ireland.)
Potatoes, to boil well together, should be all of the same sort, and
as nearly equal in size as may be. Wash off the mould, and scrub
them very clean with a hard brush, but neither scoop nor apply a
knife to them in any way, even to clear the eyes.[105] Rinse them
well, and arrange them compactly in a saucepan, so that they may
not lie loose in the water, and that a small quantity may suffice to
cover them. Pour this in cold, and when it boils, throw in about a
large teaspoonful of salt to the quart, and simmer the potatoes until
they are nearly done, but for the last two or three minutes let them
boil rapidly. When they are tender quite through, which may be
known by probing them with a fork, pour all the water from them
immediately, lift the lid of the saucepan to allow the steam to escape,
and place them on a trivet, high over the fire, or by the side of it, until
the moisture has entirely evaporated; then peel, and send them to
table as quickly as possible, either in a hot napkin, or in a dish, of
which the cover is so placed that the steam can pass off. There
should be no delay in serving them after they are once taken from
the fire. Irish families always prefer them served in their skins. Some
kinds will be sufficiently boiled in twenty minutes, others in not less
than three quarters of an hour.
105. “Because,” in the words of our clever Irish correspondent, “the water through
these parts is then admitted into the very heart of the vegetable; and the
latent heat, after cooking, is not sufficient to throw it off; this renders the
potatoes very unwholesome.”
20 minutes to 1 hour, or more.

Obs. 1.—The water in which they are boiled should barely cover
the potatoes. After it is poured off, they should be steamed for twenty
minutes or half an hour, if large.
Obs. 2.—Habitual potato-eaters know well that this vegetable is
never so good as when served in the skin the instant it is taken from
the fire, dished in a hot napkin, or sent to table without a cover over
it. It should also be clean and dry that it may at pleasure be taken in
the fingers and broken like bread, or held in the dinner napkin while
the inside is scooped out with the fork, thus forming it into a sort of
cup. The large Yorkshire Regents dressed and eaten in this way
afford in themselves an almost sufficient meal. We have found from
long daily experience, that those which averaged three, or at the
utmost four to the pound, were the best in quality, and remained so
to quite the end of their season: they required as the spring
advanced, an hour’s boiling or more.
TO BOIL POTATOES.
(The Lancashire way.)

Pare the potatoes, cover them with cold water, and boil them
slowly until they are quite tender, but watch them carefully, that they
may not be overdone; drain off the water entirely, strew some salt
over them, leave the saucepan uncovered by the side of the fire, and
shake it forcibly every minute or two, until the whole of the potatoes
appear dry and floury. Lancashire cooks dress the vegetable in this
way to perfection, but it is far from an economical mode, as a large
portion of the potato adheres to the saucepan; it has, however, many
admirers.
TO BOIL NEW POTATOES.
These are never good unless freshly dug. Take them of equal size,
and rub off the skins with a brush or a very coarse cloth, wash them
clean, and put them without salt into boiling, or at least, quite hot
water; boil them softly, and when they are tender enough to serve,
pour off the water entirely, strew some fine salt over them, give them
a shake, and let them stand by the fire in the saucepan for a minute;
then dish and serve them immediately. Some cooks throw in a small
slice of fresh butter, with the salt, and toss them gently in it after it is
dissolved. This is a good mode, but the more usual one is to send
melted butter to table with them, or to pour white sauce over them
when they are very young, and served early in the season.
Very small, 10 to 15 minutes: moderate sized, 15 to 20 minutes.
Obs.—We always, for our own eating, have new potatoes
steamed for ten minutes or longer after the water is poured from
them, and think they are much improved by the process. They
should be thoroughly boiled before this is done.
NEW POTATOES IN BUTTER.
Rub off the skins, wash the potatoes well and wipe them dry; put
them with three ounces of good butter, for a small dish, and with four
ounces or more for a large one, into a well-tinned stewpan or Keep
them well shaken or tossed, that they may be equally done, and
throw in some salt when they begin to stew. This is a good mode of
dressing them when they are very young and watery.
TO BOIL POTATOES.
(Captain Kater’s Receipt.)

Wash, wipe, and pare the potatoes, cover them with cold water,
and boil them gently until they are done, pour off the water, and
sprinkle a little fine salt over them; then take each potato separately
with a spoon, and lay it into a clean warm cloth, twist this so as to
press all the moisture from the vegetable, and render it quite round;
turn it carefully into a dish placed before the fire, throw a cloth over,
and when all are done, send them to table quickly. Potatoes dressed
in this way are mashed without the slightest trouble; it is also by far
the best method of preparing them for puddings or for cakes.
TO ROAST OR BAKE POTATOES.
Scrub and wash exceedingly clean some potatoes nearly assorted

in size; wipe them very dry, and roast them in a Dutch oven before
the fire, placing them at a distance from it, and keeping them often
turned; or arrange them in a coarse dish, and bake them in a
moderate oven. Dish them neatly in a napkin, and send them very
hot to table; serve cold butter with them. 1-3/4 to upwards of 2 hours.
SCOOPED POTATOES. (ENTREMETS.[106])
106. Or second course dish.
Wash and wipe some large

potatoes of a firm kind, and with
a small scoop adapted to the
purpose,[107] form as many
diminutive ones as will fill a dish;
cover them with cold water, and when they have boiled very gently
for five minutes pour it off, and put more cold water to them; after
they have simmered a second time for five minutes, drain the water
quite away, place the cover of the saucepan so as to leave an inch
or more of open space for the moisture to evaporate, and let them
steam by the side of the fire from four to five minutes longer. Dish
them carefully, pour white sauce over them, and serve them in the
second course. Old potatoes thus prepared, have often been made
to pass for new ones, at the best tables, at the season in which the
fresh vegetable was dearest.[108] The time required to boil them will
of course vary with their quality; we give the method which we have
found very successful.
107. This may be procured of any ironmonger.
108. Vegetables and fruit are now so generally forced and brought so early into
our markets, that there is little need of these expedients at present.
CRISPED POTATOES, OR POTATO-RIBBONS. (ENTREMETS.)
(Or to serve with Cheese.)

Wash well, and wipe, some potatoes of good flavour; cut them up
into slices of from half to a whole inch thick, free them from the skins,
and then pare them round and round in very thin, and very long
ribbons. Lay them into a pan of cold water, and half an hour before
they are wanted for table lift them on to a sieve that they may be well
drained. Fry them in good butter, which should be very hot when
they are thrown in, until they are quite crisp, and lightly browned;
drain and dry them on a soft cloth, pile them in a hot dish, strew over
them a mixed seasoning of salt and cayenne in fine powder, and
serve them without delay. For the second course, dress them in the
same manner, but omit the cayenne. Five or six minutes will fry
them.
FRIED POTATOES. (ENTREMETS.)
(A Plainer Receipt.)
After having washed them, wipe and pare some raw potatoes, cut
them in slices of equal thickness, or into thin shavings, and throw
them into plenty of boiling butter, or very pure clarified dripping. Fry
them of a fine light brown, and very crisp; lift them out with a
skimmer, drain them on a soft warm cloth, dish them very hot, and
sprinkle fine salt over them. This is an admirable way of dressing
potatoes, very common on the Continent, but less so in England
than it deserves to be. Pared in ribbons or shavings of equal width,
as in the receipt above, and served dry and well fried, lightly piled in
a dish, they make a handsome appearance, and are excellent
eating. If sliced they should be something less than a quarter of an
inch thick.
MASHED POTATOES.
Boil them perfectly tender quite through, pour off the water, and
steam them very dry by the directions already given in the receipt of
page 310, peel them quickly, take out every speck, and while they
are still hot, press the potatoes through an earthen cullender, or
bruise them to a smooth mash with a strong wooden fork or spoon,
but never pound them in a mortar, as that will reduce them to a close
heavy paste. Let them be entirely free from lumps, for nothing can be
more indicative of carelessness or want of skill on the part of the
cook, than mashed potatoes sent to table full of these. Melt in a
clean saucepan a slice of good butter with a few spoonsful of milk,
or, better still, of cream; put in the potatoes after having sprinkled
some fine salt upon them, and stir the whole over a gentle fire with a
wooden spoon, until the ingredients are well-mixed, and the whole is
very hot. It may then be served directly; or heaped high in a dish, left
rough on the surface, and browned before the fire; or it may be
pressed into a well buttered mould of handsome form, which has
been strewed with the finest bread-crumbs, and shaken free from the
loose ones, then turned out, and browned in a Dutch or common
oven. More or less liquid will be required to moisten sufficiently
potatoes of various kinds.
Potatoes mashed, 2 lbs.; salt, 1 teaspoonful; butter, 1 to 2 oz.; milk
or cream, 1/4 pint.
Obs.—Mashed potatoes are often moulded with a cup, and then
equally browned: any other shape will answer the purpose as well,
and many are of better appearance.
ENGLISH POTATO BALLS, OR CROQUETTES.
Boil some floury potatoes very dry, mash them as smoothly as

possible, season them well with salt and white pepper, warm them
with about an ounce of butter to the pound, or rather more if it will not
render them too moist, and a few spoonsful of good cream. Boil
them very dry; let them cool a little, roll them into balls, sprinkle over
them vermicelli crushed slightly with the hand, and fry them a fine
light brown. They may be dished round a shape of plain mashed
potatoes, or piled on a napkin by themselves. They may likewise be
rolled in egg and fine bread-crumbs instead of in the vermicelli, or in
ground rice, which answers very well for them.
POTATO BOULETTES. (ENTREMETS.)
(Good.)
Boil some good potatoes as dry as possible, or let them be
prepared by Captain Kater’s receipt; mash a pound of them very
smoothly, and mix with them while they are still warm, two ounces of
fresh butter, a teaspoonful of salt, a little nutmeg, the beaten and
strained yolks of four eggs, and last of all the whites thoroughly
whisked. Mould the mixture with a teaspoon and drop it into a small
pan of boiling butter, or of very pure lard, and fry the boulettes for
five minutes over a moderate fire: they should be of a fine pale
brown, and very light. Drain them well and dish them on a hot
napkin.
Potatoes, 1 lb.; butter, 2 oz.; salt, 1 teaspoonful; eggs, 4: 5
minutes.
Obs.—These boulettes are exceeding light and delicate, and make
an excellent dish for the second course; but we think that a few
spoonsful of sweet fresh cream boiled with them until the mixture
becomes dry, would both enrich them and improve their flavour. They
should be dropped into the pan with the teaspoon, as they ought to
be small, and they will swell in the cooking.
POTATO RISSOLES.
(French.)
Mash and season the potatoes with salt, and white pepper or
cayenne, and mix with them plenty of minced parsley, and a small
quantity of green onions, or eschalots; add sufficient yolks of eggs to
bind the mixture together, roll it into small balls, and fry them in
plenty of lard or butter over a moderate fire, or they will be too much
browned before they are done through. Ham, or any other kind of
meat finely minced, may be substituted for the herbs, or added to
them.
POTATOES À LA MAÎTRE D’HÔTEL.
Boil in the usual manner some potatoes of a firm kind, peel, and
let them cool; then cut them equally into quarter-inch slices. Dissolve
in a very clean stewpan or saucepan from two to four ounces of
good butter, stir to it a small dessertspoonful of flour, and shake the
pan over the fire for two or three minutes; add by slow degrees a
small cupful of boiling water, some pepper, salt, and a tablespoonful
of minced parsley; put in the potatoes, and toss them gently over a
clear fire until they are quite hot, and the sauce adheres well to
them: at the instant of serving add a dessertspoonful of strained
lemon-juice. Pale veal gravy may be substituted for the water; and
the potatoes after being thickly sliced, may be quickly cut of the
same size with a small round cutter.
POTATOES À LA CRÈME.
Prepare the potatoes as above, and toss them gently in a quarter

of a pint or more of thick white sauce or of common bechamel, with
or without the addition of the minced parsley.

Energy Transition, Climate Change, and COVID-19: Economic Impacts of The Pandemic 1st Edition Fateh Belaïd All Chapter Instant Download

Uploaded by

Copyright:

Available Formats

Energy Transition, Climate Change, and COVID-19: Economic Impacts of The Pandemic 1st Edition Fateh Belaïd All Chapter Instant Download

Uploaded by

Document Information

Original Title

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

Energy Transition, Climate Change, and COVID-19: Economic Impacts of The Pandemic 1st Edition Fateh Belaïd All Chapter Instant Download

Uploaded by

Copyright:

Available Formats

Full download test bank at ebookmeta.

Energy Transition, Climate Change, and COVID-19:

Download More ebooks from https://ebookmeta.com

Exploring the Consequences of the Covid-19 Pandemic:

Economic Recovery After COVID-19

Cybersecurity in the COVID 19 Pandemic 1st Edition

The Unequal Pandemic COVID 19 and Health Inequalities

COVID-19 Pandemic: Lessons from the Frontline Jorge

Power, Media and the Covid-19 Pandemic: Framing Public

Stress Tested The COVID 19 Pandemic and Canadian

Pandemics and Travel COVID 19 Impacts in the Tourism

Energy Transition, Climate

ISBN 978-3-030-79712-6 ISBN 978-3-030-79713-3 (eBook)

© Springer Nature Switzerland AG 2021

The impact of COVID-19 pandemic is far beyond health implications. No other

Various COVID-19 impacts have been discussed: sustainability issues, potential

Lille, France Fateh Belaïd

Paris, France Anna Cretì

Assessing the Relationship Between Air Quality, Wealth,

A Proposition Relationship Between Green Workplace

Roberto Antonietti, Paolo Falbo, and Fulvio Fontini

1 Data obtained from the World Health Organization website: https://www.who.int.

R. Antonietti · F. Fontini ()

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 1

2 Data and Preliminary Analysis

by the ECDC’s Epidemic Intelligence team, based on reports provided by national

Table 1 Summary statistics

2 For more information, see: https://www.ecdc.europa.eu/en/covid-19/data-collection.

variable because it is collinear with AGRVA/GDP and MANVA/GDP.

Table 3 Infections and deaths: relationship with pollution and wealth

Table 4 Summary statistics

Fig. 1 Clusters of countries

In this chapter, we have analyzed the relationship between pollution, measured

Table 6 Clusters of countries

Table 7 Eigenvectors of the Cluster Eigenvalue Difference Proportion Cumulative

Table 8 Eigenvalues of the correlation matrix

Table 9 Average values of clusters

Côme Billard and Anna Creti

C. Billard () · A. Creti

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 15

warming. See full description in the Sect. 2.

10 Inthe following, we name “cascading process” such a dynamics of emission contraction.

as manufacture of other non-metallic mineral products20 are highly present in the

on the examined country.

2 Methodology and Data

2.1 The Emission Reduction Multiplier Matrix

Following Cahen-Fourot et al. (2020), we start with the national inter-industry

Fig. 1 A stylised input–output table (Cahen-Fourot et al., 2020)

24 See Miller and Blair, 2009; Cahen-Fourot et al., 2020.

We then transpose G to be able to read the effects of changes in sectoral primary

26 Augustinovics, 1970; Beyers, 1976.

Overall, this methodological approach allows us to investigate both internal

2.2 Datasets: Input–Output Tables and Emissions

We apply the methodology described above to five European economies, heteroge-

Table 1 Breakdown of examined NACE sectors

29 OECD Statistics: https://stats.oecd.org/Index.aspx?DataSetCode=IOTSI4_2018.

NACE, is the classification of economic activities in the European Union (EU).

3 Cascading Greenhouse Gas Emissions

3.1 Emission Coefficients