Applied Energy 295 (2021) 117051

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

The impact of temperature increase on firm profitability. Empirical

evidence from the European energy and gas sectors
Sorin Gabriel Anton *
“Alexandru Ioan Cuza” University of Ias¸i, Faculty of Economics and Business Administration, Ias¸i, Romania

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• The impact of temperature change on

firm profitability is analyzed.
• A quantile regression approach is
employed on firms active in the energy
and gas sector.
• An increase in the annual temperature
has a positive impact on firm
profitability.
• This positive effect varies at different
quantiles of the performance
distribution.
• Market concentration and energy prices
have a significant influence on firm
profitability.

A R T I C L E I N F O
A B S T R A C T
Keywords:
Although temperature change is affecting economic activity, there are a few empirical studies on how it in-
Climate change
Firm profitability
fluences firm performance. Using panel data of 147 European listed firms active in the energy and gas sectors
Quantile regression over the period 2009–2016, this study aims to analyze the impact of temperature change on firm profitability
Energy sector while controlling for firm-specific and energy market-related factors. The results of the quantile regression
Europe approach show that an increase in the annual temperature has a positive impact on firm profitability.
However, this impact varies at different quantiles of the profitability distribution. A 1℃ degree increase in
temperature is related to a 1.30% increase in operating profitability for the most profitable firms (Q90).
However, the
magnitude of this impact is lower for less profitable firms (Q10 and Q25). The empirical results revealed that
cash flows, market opportunities, leverage, firm size, net asset turnover, as well as market concentration and
energy prices, have a significant influence on firm profitability in the energy and gas sectors. These findings
appear to be robust to several control variables and robustness tests. The study provides empirical evidence on
the impact of climate change on the performance of firms active in climate-sensitive sectors (such as energy
and gas sectors). The findings of this study are particularly helpful for portfolio risk managers, energy traders,
and policymakers.

* Address: Carol I Avenue, No. 11, 700505 Iasi, Romania.

E-mail address: [email protected].

https://doi.org/10.1016/j.apenergy.2021.117051
Received 1 January 2021; Received in revised form 23 April 2021; Accepted 30 April 2021
Available online 15 May 2021
0306-2619/© 2021 Elsevier Ltd. All rights reserved.
S.G. Anton Applied Energy 295 (2021) 117051

1. Introduction
this study. Section 4 presents the empirical results and several robust-
ness tests. The last section presents the conclusions and discusses the
An increasing number of academic studies and official reports pro-
implications of our findings.
vide empirical evidence on climate change. Several studies document
an increase in mean temperatures in the last decades [1,2,3,4] and a
2. Literature review and hypothesis development
higher probability of extreme weather events [3]. Brabazon and Idowu
[5] highlight that 70% of world economic activity is susceptible to
In the last decade, several empirical papers have examined the
weather influences. Despite these evolutions and their significant socio-
impact of rising temperature on macroeconomic variables
economic impact, we know little about how weather and rising
[8,9,10,11,12,13] or several sectors [14]. For example, Dell et al. [13]
temperatures affect the performance of the (European) firms [6].
and Hsiang [11] find a negative relationship between an increase in
Furthermore, there is no empirical study on the impact of temperature
mean temperature and per capita income. Employing a large sample of
increase on the profit- ability of European firms in the energy and gas
countries, Dell et al. [13] find that a 1℃ increase in mean temperature
sectors even if they play a key role in any economy. Additional empirical
is
evidence on this impact is necessary to support decision-making or
related to a 1.4% decrease in per capita income in developing
public policies on climate change in the energy and gas sectors.
countries. Using a large sample of countries, Jones and Olken [10]
In order to cover this gap in the extant literature, the paper aims to
show that temperature shocks have a negative impact on exports.
assess the impact of climate change on firm profitability using a
Bansal and Ochoa [12] highlight that an increase in global temperature
sample of 147 listed firms located in 21 countries from the European Union
harms economic growth in countries closer to the Equator.
(EU) and active in climate-sensitive sectors – energy and gas
Recently, a growing number of studies focus on the impact of
production and transportation. Understanding the potential impact of
rising temperature (climate change) on firm-level variables
temperature in- crease on these sectors is particularly relevant for the
(microeconomic level), studying the impact of climate/temperature
EU countries for three reasons. Firstly, the EU countries are highly
change on: stock market returns [15,16]; sales, productivity, and
dependent on energy for the growth of their national economies.
profitability [3,17,2,18]; the cost of equity [19]; the systemic risk [20].
Secondly, the EU framework is interesting to study given the high
However, the empirical results are rather mixed, still scarce, and
commitment to climate change mitigation in its member countries and
focused only on a few countries (mainly the U.S.). Addoum et al. [3]
various policies implemented at the EU and national level. Thirdly,
find no statistically sig- nificant effect of temperature exposure on
energy and gas production and transportation are highly exposed to
annual and quarterly firm- level sales in the case of U.S. firms over
weather-related risks [7].
the period 1990-2015, excepting the energy sector. Also, they found
Employing annual data on over 147 European listed firms over the
no impact of tempera- ture change on productivity and profitability.
period 2009 to 2016, the paper shows that (1) the temperature increase
Recenlty, Jawid [17] employed a sample of 1,502 farmers from 14
has a positive impact on firm profitability and (2) this impact is
districts in Afganistan over the period 1979 to 2014 to study the
different for the quantiles of the profitability distribution. This finding
impact of temperature and pre- cipitation change on net crop
is robust to several sensitivity tests, including alternative measures of
revenue. The results show a positive relationship between increased
profitability and firm size. Also, the econometric analysis shows that
annual temperature and net crop reve- nue. Employing a sample of
cash flows, market opportunities, leverage, firm size, net asset
600 listed firms from the European Union, Tzouvanas et al. [20]
turnover, as well as market concentration and energy prices, are key
provide empirical evidence for a nonlinear effect of temperature on
determinants of firm profitability in the energy and gas sectors.
systemic risk. Balvers et al. [19] demonstrate the negative impact of
This study has several important contributions to the extant liter-
temperature shocks on the average cost of equity capital on a
ature on firm profitability and climate change. Firstly, to the best of the
sample of U.S. firms for the time span 1953 to 2015.
author’s knowledge, this is the first study to assess the impact of tem-
An increasing number of studies have analyzed the impact of
perature increase on firm profitability focusing on energy and gas sec- climate
tors from the European Union. Thus, the paper contributes to the change on the energy sector but they focus only on production [21],
growing literature on the impact of climate change at the microeco- transmission (infrastructure2), and demand [22,23]. However, the
nomic level. Secondly, in terms of methodology, the study employs a relationship between changing (rising) temperature and firm sales in
quantile regression (QR) approach 1 that has several significant advan- the energy and gas sectors is unclear3 although air temperature has a
tages over the mean approaches employed in the previous literature. strong impact on sales and this impact is almost immediate. On the one
The methodology employed in the study offers more information on hand, some papers documented a negative relationship between
the relationship between temperature increase and firm profitability, temperature and energy demand - the lower the outside temperature,
sug- gesting that the effect is not homogeneous as reported in previous the higher the amount of energy consumption [24]. Addoum et al. [3]
research [3]. Thirdly, as the empirical analysis is focused on firms provide empirical evidence for a population average effect in the case
active in energy and gas sectors, the employed approach controls for of energy firms employing a sample of U.S. firms over the period 1990-
the effect of energy market characteristics and firm-level variables on 2015. More exactly, the authors found that “an abnormal day of
firm prof- itability. Our results highlight that energy market-related exposure to extreme cold is associated with about a 1.28- to 1.31-
factors are important determinants of firm profitability but their effects percentage-point increase in sales” [3]. On the one hand, extreme
depend on the profitability distribution. Fourthly, contrary to the extant temperatures (both hot and cold temperatures) are associated with
literature, this paper uses several proxies for firm profitability and size greater energy demand [25] from households and industries, and thus
in order to test the robustness of the main findings. firms’ sales are increasing.
The remainder of this study is organized as follows. Section 2 pro- The relationship between temperature increase and firm
vides a brief literature review and introduces the hypothesis. Section 3 profitability has been analyzed only in the case of U.S. firms [2,3] but
describes the sample selection and the empirical methods employed in the results are rather mixed and inconclusive. To the best of the
author’s knowledge, there is no study on the impact of climate change
on the profitability of European listed firms in the energy and gas
sectors despite the economic and strategic importance of these firms.
This is the gap that I will like to address in this empirical paper.

1
The initial descriptive statistics (skewness and kurtosis) show that the dis-
tributions of the profitability measures are non-Gaussian, implying that the general regression methods (i.e. OLS or fixed-effect estimator) provide “coef-
2
S.G. Anton Applied Energy 295 (2021) 117051
ficient estimates that are unrepresentative of the overall performance distri- 2
See Cronin et al. [26] for a detailed review on the impacts of climate
bution” [31]. change on the energy systems.
3
For example, higher temperature will increase energy demand for cooling
and decrease fuel and energy demand for heating.

3
S.G. Anton Applied Energy 295 (2021) 117051

Table 1
Finland, France, Latvia, Lithuania, Germany, Greece, Hungary, Italy,
Variables employed in the study.
Luxembourg, Poland, Portugal, Romania, the Netherlands, Spain,
Role/Variable (symbol) Definition/Computation Data
Swe- den, and the United Kingdom) over the period 2009-2016 5. Firms
source
included in our sample are active in the production of electricity
Dependent variables
(NACE 35.11, Rev. 2.0), the transmission of electricity (NACE 35.12,
Rev. 2.0),
Operating profit margin or (EBIT/Operating revenue)*100 Orbis the distribution of electricity (NACE 35.13, Rev. 2.0), the distribution
EBIT margin (EBMA)
of gaseous fuels through mains (NACE 35.22, Rev. 2.0), and the trade of
gas
Return on equity (ROE) (Profit before tax/Shareholders funds) Orbis
through mains (NACE 35.23, Rev. 2.0). In line with extant empirical
*100
Independent variables
literature (Jarait´e and Kaˇzukauskas, 2013), unconsolidated financial
Explanatory variables accounts are employed in the empirical analysis.
Temperature (TEMP) Annual temperature data corresponds World
to Bank
the average of monthly observations 3.2. Variable selection
(Celsius) World
Rainfall (RAIN) Annual rainfall data corresponds to the Bank In line with extant literature [29,30], our dependent variable is
average of monthly observations (MM)
Control variables
operating profit margin or EBIT margin (EWBA) as a common
Cash flows (CFTA) Cash Flows/Total assets Orbis
Tobin’s Q (TQ) Market capitalization/Total assets Orbis accounting-earnings-based performance measure6. EWBA is determined
Size (SIZE) The natural logarithm of total assets Orbis as the ratio between earnings before net interest and tax (EBIT) and
book value (in the correlation and
operating revenues (See Table 1). The study investigates the changes
regression analyses)
Size (SIZE1) The natural logarithm of enterprise value in the operating performance caused by the variations in the
Orbis
(market capitalization) (in the temperature and thus the focus is on the actual direct impact of
correlation and regression analyses) temperature on operating performance rather than the investors’
Leverage (LEV) (Non-current liabilities + Loans)/Total Orbis expectations of this effect on market capitalization. Also, EBIT provides
assets
a better comparison of operating performance across countries as it
takes out the impact of
Current ratio (CURR) Current assets/Current liabilities Orbis interest and taxes on profitability. In the second part of our analysis, I
Net asset turnover (NAR) Operating revenue/(Shareholders funds Orbis
+ Non-current liabilities) redo the tests with return on equity (ROE) as the proxy variable for
Electricity price (EP) The price of electrical energy (without Eurostat firm profitability to check the robustness of the results. Return on
taxes) for industrial customers equity, computed as profit before tax over shareholders’ funds, has
also been
Market concentration The market share of the largest generator Eurostat
used as an accounting-based measure of performance [31].
(MC) in the electricity market (% of total
electricity generation)
In line with previous empirical papers [12], I employed annual data
on temperature and rainfall covering the period 2009–2016 obtained
from the Climate Change Knowledge Portal [27]. Annual temperature
Addoum et al. [3] found a positive relationship between an increase and rainfall data are computed as the average of monthly observations.
in mean temperature and net income in the case of U.S. energy firms As highlighted in the extant literature, numerous factors affect firm
but “the statistical strength of the result is only marginal” and the profitability [30,32]. For the purpose of this paper and given the
economic effect small. Hugon and Law [2] employ a large sample of limited data availability, I focus only on firm-level variables and on
U.S. firms comprising 159,584 firm-year observations for a long time character- istics of electricity markets. Following the extant literature
span (1955- [33,32], I include as control variables at the firm level the following:
2016) to analyze the effect of climate change on earnings. Their results cash-flows scaled by the book value of total assets (CFTA), market
highlight that a 1℃ increase in mean temperature is related to a 1.6% opportunities proxied by Tobin’s Q (TQ), leverage (LEV), firm size (SIZE)
decrease in earnings. Furthermore, for less diversified firms, the nega- measured as the natural logarithm of total assets book value (and
tive impact is higher - 2.5% decrease in sales. However, some firms alternatively as the natural logarithm of market capitalization), net
benefit from warmer climate – the winners represent close to 50% in asset turnover (NAR),
healthcare (52.6%) and energy, oil, and gas (46.7%) sectors. Based on and liquidity (CURR) measured by the current ratio.
these studies, we test the following research hypothesis: Furthermore, since the characteristics of energy markets affect a
firm’s performance [30], the price of electrical energy (without taxes)
H1: There is a positive relationship between temperature increase for industrial customers and the market share of the largest generator
and firm profitability for firms in the energy and gas sectors. in the electricity market have been employed in the empirical models.

3. Data and empirical strategy

3.3. Econometric specifications
3.1. Sample description
I conduct the analysis in several steps. Following the extant
The firm-level financial data comes from the Orbis [26] database literature [34,3], in the first part of the empirical analysis, I do not
(developed and maintained by Bureau Van Dijk) while data on tem- include another firm- or energy market-level control variables in the
perature and rainfall is taken from the Climate Change Knowledge models to avoid the “overcontrolling” issues. Thus, the following
Portal [27]. Energy market-related data (electricity price and market panel model is estimated:
concentration) has been retrieved from Eurostat [28]. EBMAi,j,t = αi,j + β1 TEMPj,t + β2 RAINj,t + εi,j,t (1)
After usual cleaning procedures have been applied4, the final
sample comprises 774 firm-year observations for 147 listed firms from where EBMAi,j,t is the operating profitability of firm i, located in
21 Eu- ropean countries (Austria, Belgium, Bulgaria, Czech Republic, country j, at time t; Tempj,t represents the annual temperature
Denmark, computed as the average of monthly observations in country j
over year t; Rainj,t
4
S.G. Anton Applied Energy 295 (2021) 117051

5
The sample period ended in 2016 since temperature and rainfall data
4
I eliminate observations with negative total assets, I drop firms with (provided by World Bank using the same methodology) is available until his
incomplete records on main regression variables and firms with less than 3 year.
years of consecutive observations. The final unbalanced panel data helps to 6
See Lee and Li [29] and Li et al. [31] for a detailed description of the ad-
mitigate the survivorship bias problem [47]. vantages of the accounting-based performance indicators.

5
S.G. Anton Applied Energy 295 (2021) 117051

Table 2
Descriptive statistics.
Variable Observations Mean Std. Dev. 25th percentile 50th percentile 75th percentile Skewness Kurtosis

EBMA 774 11.46 17.55 3.00 10.18 20.53 —0.62 6.49

ROE 774 2.98 35.00 0.17 8.72 15.52 —3.06 17.36
TEMP 774 10.47 2.69 9.02 9.98 12.50 —0.53 4.10
RAIN 774 68.78 19.69 54.67 63.85 78.36 0.94 3.12
CFTA 774 0.04 0.13 0.03 0.06 0.08 —3.94 23.74
TQ 774 0.51 0.62 0.21 0.36 0.61 4.94 35.00
LEV 774 0.42 0.22 0.24 0.45 0.58 —0.11 2.27
SIZE 774 13.93 2.71 12.02 13.90 15.91 —0.18 2.70
SIZE1 763 13.47 2.67 11.78 13.23 15.62 —0.20 2.45
NAT 774 0.93 1.69 0.21 0.54 0.93 5.04 31.34
CURR 774 1.94 3.67 0.89 1.17 1.77 7.33 62.02
MC 774 40.10 21.36 24.50 32.00 51.70 1.01 2.86
EP 774 0.11 0.06 0.08 0.09 0.11 2.16 6.45

Notes: EBMA stands for operating profit margin; ROE - return on equity; TEMP – temperature; RAIN – rainfall; CFTA - cash flows; TQ - Tobin’s Q; SIZE – size measured
as log of total assets; SIZE1 - size measured as log of enterprise value; LEV - leverage; CURR - current ratio; NAR - net asset turnover; EP - electricity price; MC - market
concentration.

measures the amount of rainfall over the year t in country j; εi,j,t is an

error term. with
In the second part of our analysis, I extend model 1 by adding firm )
Quantθ yi,t|xi,t = ix’ tβθ (4)
and energy market-level characteristics that affect the operating prof- ,
itability in the energy and gas sectors [30]. with the following restriction:
The following linear regression model for panel data is estimated: )
Quantθ uθi,t|xi,t = 0 (5)
EBMAi,j,t = αi,j + β1TEMPj,t + β2RAINj,t + β3FIRM LEVi,j,t
+ β4EN MARKET LEVj,t + εi,j,t where yi,t represent the dependent variable (firm profitability), xi,t
(2)
represent a vector of regressors, βθ is the vector of parameters to be
where the vector FIRM_LEV
includes
namely firm-specific variables, estimated, uθi,t is the error, i denotes firm (i = 1,., 147) and t denotes
time
i,j,t (
cash flows, Tobin’s Q, firm size, leverage, current ratio, and net asset (t = 1,., 8). Quantθ yi,t |xi,t denotes the θth conditional quantile of yi,t
)

turnover; the vector EN_Market_LEVj,t captures variables that given xi,t. Standard errors (SE) for the coefficients are based on the
describe the electricity markets such as the price of electrical energy for bootstrap method with 1,000 replications as this approach is robust for
industrial customers and market concentration for country j at time t; small samples [39], it is valid under many forms of heterogeneity [40]
β1, …, β4 are the coefficients to be estimated; εi,j,t is an error term. and it provides more suitable estimations for SE than the asymptotic
method [41].
3.4. Methodology
4. Empirical results and discussion
The quantile regression (hereafter, QR) is employed in the paper
4.1. Summary statistics and correlation
because it offers more information about the relationship between
temperature increase and firm operating performance. Furthermore,
this econometric approach is robust to the existence of outliers [35] The descriptive statistics of the variables employed in the empirical
and the non-Gaussian error distribution [36]. Given these advantages, analysis are shown in Table 2. All independent financial variables
an increasing number of studies [29,31,33,37] have employed QR were winsorized at the 1st and 99th percentiles in order to overcome the
meth- odology in the last years in the corporate finance literature. in- fluence of outliers before presenting the summary statistics. It is
Koenker and Bassett [38], in their seminal paper, developed the QR noticeable that the average EBMA is 11.46% with the mean being
approach with the following specification: 10.18%. Operating firm profitability is skewed to the right, meaning it
has long right tails. The second measure of firm profitability (ROE) has a
’
yi,t = xi, βθ + uθi,t (3) mean of 2.98% and a median of 8.72%. The distribution is negatively
t
Temperature

6
 4
ES
2
FI
0
2009 2010 2011 2012 2013 2014 2015 2016 FR
S.G. Anton Applied Energy 295 (2021) 117051
Year GB
Table 2

Fig. 1. Evolution of annual temperature in selected EU countries: 2009 to 2016.

7
S.G. Anton Applied Energy 295 (2021) 117051

Table 3
Correlation matrix.
EBMA ROE TEMP RAIN CFTA TQ LEV SIZE SIZE1 NAT CURR MC EP

EBMA 1
ROE 0.3978 1
TEMP 0.1499 0.0611 1
RAIN —0.006 —0.1202 —0.0567 1
CFTA 0.3945 0.577 0.0476 —0.185 1
TQ —0.0353 —0.0644 —0.1871 0.1179 —0.1549 1
LEV 0.229 0.0457 0.2678 0.1073 0.0114 —0.3221 1
SIZE 0.2504 0.2542 0.2258 —0.0123 0.3037 —0.332 0.3595 1
SIZE1 0.2944 0.2673 0.2212 0.0259 0.3201 —0.1658 0.4059 0.9683 1
NAT —0.2174 —0.0085 —0.0293 —0.0747 —0.0207 0.0396 —0.1584 —0.1479 —0.1669 1
CURR —0.0963 —0.0619 —0.0912 0.069 —0.1069 0.1505 —0.3157 —0.1533 —0.1727 —0.0993 1
MC —0.0399 —0.01 0.1142 0.1921 0.0081 —0.032 0.0843 —0.0035 —0.0012 —0.0343 0.06 1
EP —0.0652 —0.012 —0.0103 —0.0768 0.0003 —0.1194 0.2327 0.1603 0.1424 —0.0055 —0.0308 —0.215 1

Notes: EBMA stands for operating profit margin; ROE - return on equity; TEMP – temperature; RAIN – rainfall; CFTA - cash flows; TQ - Tobin’s Q; SIZE – size measured
as log of total assets; SIZE1 - size measured as log of enterprise value; LEV - leverage; CURR - current ratio; NAR - net asset turnover; EP - electricity price; MC - market
concentration.

Table 4 Energy prices are, on average, the highest in Germany, while the
Variance inflation factor (VIF) statistics of explanatory variables. lowest are found in Denmark and Bulgaria.
Table 3 contains the correlation coefficients between the dependent
Variable VIF 1/VIF
and independent variables. A positive correlation between temperature
TEMP 1.13 0.88
RAIN 1.14 0.88
and operating profitability/return on equity can be observed. The
CFTA 1.17 0.85 cor- relation coefficients are below 41%7 suggesting that
TQ 1.21 0.83 multicollinearity is unlikely to be an issue in the models.
LEV 1.54 0.65 Furthermore, in order to check the possible existence of collinearity
SIZE 1.42 0.70 between explanatory variables, the variance inflation factor (VIF) sta-
NAT 1.08 0.93
tistic was calculated and reported in Table 4. As all VIF values were
CURR 1.19 0.84
MC 1.13 0.88 below the threshold of 10.00 [42], I consider that the presence of mul-
EP 1.15 0.87 ticollinearity can be dismissed.
Also, I employ a Fisher-type unit-root test appropriate for the un-
Notes: TEMP – stands for temperature; RAIN – rainfall; CFTA - cash flows; TQ
- Tobin’s Q; SIZE – size measured as log of total assets; LEV - leverage; CURR
balanced panel to investigate the series stationarity. The results (not
- current ratio; NAR - net asset turnover; EP - electricity price; MC - market presented here, but available upon request) strongly suggest that all the
concentration. series are stationary processes at the conventional levels.

Table 5
Simple regression.
OLS Q10 Q25 Q50 Q75 Q90

TEMP 0.9486* 0.3176 0.8878*** 0.6556*** 1.0116** 1.2228**

(0.5251) (0.4306) (0.1794) (0.1272) (0.4469) (0.5603)
RAIN 0.0071 0.0468 —0.0287 —0.0097 0.1086 —0.0348
(0.0563) (0.0524) (0.0231) (0.0223) (0.0673) (0.0622)
Constant 1.0434 —10.3523* —4.3646 3.8857* 2.2989 23.6207***
(6.4274) (6.2388) (2.6965) (2.2400) (5.8408) (7.6445)
R2/Pseudo R2 0.0211 0.0039 0.0159 0.0120 0.0135 0.0259
Observations 774 774 774 774 774 774

Notes: The dependent variable is firm profitability (EBWA). ***, ** and * stand for significance levels 1%, 5% and * at 10%. The numbers in parentheses are robust
SEs in column (1) and bootstrapped standard errors (BSE) in columns (2)–(6). TEMP stands for temperature; RAIN – rainfall.

skewed and thus different from the normal distribution. The distribu-
4.2. Empirical results
tions of firm profitability raise the efficiency of quantile regression.
Since this paper focuses on temperature increase, it is interesting to
4.2.1. Univariate analysis
see how this has changed over time for the 21 EU countries included
First, I present the simple relationship between temperature and
in the empirical analysis. The average annual temperature has
firm profitability across the distribution. The results of the univariate
increased over the sample period from a low of 8.8 ℃ in 2010 to a
high of 10.54 ℃ anal- ysis presented in Table 5 show that the impact of increased
in 2014. However, there is a higher heterogeneity across space and temperature on operating profit margin is not homogeneous among
time, various quantile levels. For low quantiles (Q10), the effect of increased
as depicted in figure 1. Also a considerable variation in the quantity of temperature on firm profitability is positive but statistically
precipitation across countries in the sample can be observed. insignificant. However, for middle-range and high quantiles (i.e.,
Regarding firm-level characteristics, the analyzed firms, on from Q25 to Q90) temperature
average, are large (SIZE = 13.93, SIZE1 =13.47), with a moderate
leverage (LEV
= 0.42). The average Tobin’s Q is approximately 0.51 with a median
of
0.36. Concerning energy markets’ variables, it is noticeable a high
heterogeneity for both variables. The highest level of market concen- tration is found in Greece (2009), while the lowest in Lithuania (2016).
8
S.G. Anton Applied Energy 295 (2021) 117051

Table 3
7
Only correlation between the two measures of size is very high, but this
is not an issue as these two variables are employed interchangeble in the
baseline regression and robustness checks.

9
S.G. Anton Applied Energy 295 (2021) 117051

Table 10
The impact of temperature on firm profitability.
OLS Q10 Q25 Q50 Q75 Q90

TEMP 0.5807 0.8739** 0.4717*** 0.3152 0.3937 1.3022***

(0.3791) (0.4208) (0.1585) (0.1942) (0.3839) (0.2570)
RAIN 0.0293 0.0255 0.0203 0.0303 —0.0247 —0.0778*
(0.0476) (0.0447) (0.0223) (0.0224) (0.0415) (0.0426)
CFTA 52.4022*** 63.2855*** 60.6643*** 60.7553*** 37.4683*** 31.0701**
(9.7651) (21.1364) (11.6477) (12.8008) (13.2849) (13.4455)
TQ 2.9491* —0.3697 2.0308 2.7216 3.5010 4.7283*
(1.5649) (2.8799) (1.5451) (1.6636) (2.2264) (2.6812)
LEV 17.2244*** 2.2555 4.7510 18.6650*** 35.1432*** 36.9660***
(6.2499) (6.6615) (3.8216) (3.9433) (4.3296) (5.2407)
SIZE 0.4539 1.6528*** 0.6363*** —0.1778 —0.5629 —0.9237
(0.3869) (0.4172) (0.2007) (0.2472) (0.3807) (0.6505)
NAT —1.6580*** —0.0983 —1.8696** —1.7126** —1.5910*** —1.9725***
(0.5901) (1.3003) (0.8495) (0.6923) (0.4722) (0.4133)
CURR 0.0509 —0.0877 —0.1773 0.0643 0.4703 0.7349
(0.3128) (0.8398) (0.2848) (0.3580) (0.4199) (0.6849)
MC —0.0862 —0.1757*** —0.0528* —0.0148 0.0029 —0.1274**
(0.0535) (0.0644) (0.0310) (0.0215) (0.0348) (0.0589)
EP —38.8723*** —34.6914** —21.2907*** —31.8115*** —56.2149*** —95.4513***
(14.2389) (13.9508) (6.1225) (5.5061) (10.7076) (12.9671)
Constant —4.3548 —29.1071*** —9.2043** 0.9909 13.3799* 31.2356**
(6.0936) (7.3322) (3.5825) (4.0303) (7.3557) (13.8760)
R2/Pseudo R2 0.2720 0.2155 0.1807 0.1621 0.1946 0.2151
Observations 774 774 774 774 774 774

Notes: The dependent variable is firm profitability (EBWA). ***, ** and * stand for significance levels 1%, 5% and * at 10%. The numbers in parentheses are robust
SEs in column (1) and BSEs in columns (2)–(6). EBMA stands for operating profit margin; TEMP – temperature; RAIN – rainfall; CFTA - cash flows; TQ - Tobin’s
Q; SIZE – size measured as log of total assets; LEV - leverage; CURR - current ratio; NAR - net asset turnover; EP - electricity price; MC - market concentration.

Fig. 2. Variation in the “Temp, Rain, SIZE, MC, and EP” coefficient over the conditional quantiles. Notes: Horizontal lines represent OLS estimates with a 95%
confidence level. Graphs have been made using the “grqreg” Stata module.

10
S.G. Anton Applied Energy 295 (2021) 117051

Table 11
The impact of temperature on firm profitability: alternative measure Return on Equity.
OLS Q10 Q25 Q50 Q75 Q90

TEMP 0.2137 0.6680 0.6775*** 0.4752*** 0.4059** 0.4922*

(0.3805) (0.6272) (0.2491) (0.1339) (0.1587) (0.2635)
RAIN —0.0353 —0.0924 0.0857*** 0.0450** 0.0260 0.0410
(0.0699) (0.1200) (0.0266) (0.0218) (0.0273) (0.0560)
CFTA 164.0561*** 221.1581*** 215.0497*** 175.0557*** 153.0911*** 117.2787***
(31.9763) (15.2806) (16.9095) (16.6642) (17.8537) (29.4978)
TQ 3.7649 5.3408 5.7920*** 5.1204*** 7.5979*** 11.4840**
(3.2989) (5.6731) (2.0421) (1.5917) (2.1876) (4.9979)
LEV 7.9229 —7.1313 6.3470 14.4279*** 26.7866*** 40.6851***
(8.8334) (8.7851) (4.4655) (3.0096) (3.0894) (4.9713)
SIZE 1.2261* 2.2453*** 1.0813*** 0.4274** —0.2509 —1.4343***
(0.6981) (0.7769) (0.2769) (0.1730) (0.2730) (0.4061)
NAT 0.6793 —7.7081** —0.6369 0.9630 2.4600*** 6.7185***
(0.8685) (3.2216) (1.1234) (0.6240) (0.9028) (2.2712)
CURR 0.2855 0.0665 0.1970 0.2246* 0.2542** 0.0722
(0.2050) (0.1831) (0.1505) (0.1220) (0.1153) (0.0918)
MC —0.0443 —0.0084 0.0087 0.0040 0.0001 0.0227
(0.0719) (0.0768) (0.0238) (0.0149) (0.0215) (0.0362)
EP —19.2527 –32.7613 11.4945 7.6537 13.3042 34.9522
(22.4943) (50.7727) (19.3881) (7.3312) (9.6975) (24.3400)
Constant –22.7108* —44.3348*** —44.8137*** —26.5284*** —15.5803*** —3.9120
(12.6644) (13.7549) (6.7809) (3.9176) (4.1840) (6.7986)
R2/Pseudo R2 0.3856 0.3871 0.3343 0.2863 0.2489 0.2532
Observations 774 774 774 774 774 774

Notes: The dependent variable is firm profitability (ROE). ***, ** and * stand for significance levels 1%, 5% and * at 10%. The numbers in parentheses are robust SEs
in column (1) and BSEs in columns (2)–(6). ROE stands for return on equity; TEMP – temperature; RAIN – rainfall; CFTA - cash flows; TQ - Tobin’s Q; SIZE – size
measured as log of total assets; LEV - leverage; CURR - current ratio; NAR - net asset turnover; EP - electricity price; MC - market concentration.

has a significantly positive impact on firm profitability and the magni-

reported by other empirical studies focused on listed firms. For
tude of this impact is larger for higher quantiles (Q75-Q90).
example, the level of cash flows has a significant positive impact on
In the second part of the analysis, I include in equation (2) firm and
firm profit- ability at every quantile level. However, its improvement
energy market-level characteristics. Table 6 reports the estimates of
effects are stronger at lower quantiles. Market opportunities (proxied
the quantile regressions8 results (columns 2-6) along with ordinary
by Tobin’s Q
least squares (OLS) regression results (column 1). I introduce mean
- TQ) have a positive impact on firm profitability but only for high
regression (OLS) only as a benchmark as it focuses on the central
quantiles (Q90).
tendency of the distribution. A major drawback of this approach is that
Contrary to the previous research on listed firms 10, a positive rela-
it does not reveal the impact of explanatory variables on firm
tionship between the degree of indebtedness and firm profitability for
profitability for low and high profitable firms. I consider that OLS
middle and high quantiles (Q50-Q90) has been found, which rejects
estimation is quite misleading as it predicts a positive but statistically
the Pecking Order Theory of capital structure. One explanation for
insignificant relationship be- tween temperature and firm profitability.
these results is that firms in these sectors take a lot of debt to finance
The empirical results suggest that the impact of increased tempera-
their expensive investments and efficient investments contribute to
ture on firm operating profit margin is not homogeneous among
higher profitability.
various quantile levels (see also Fig. 2 for a graphical representation).
In line with previous research [30], a positive relationship between
Thus, the temperature has a significantly positive effect on firm
firm size and firm profitability has been found but only for low
profitability only for low (Q10 – Q25) and high quantiles (Q90). The
quantiles (q10-q25). These results suggest that larger firms in the
effect of increased temperature on firm profitability is insignificant for
energy and gas sectors earn higher profits.
firms at the middle- range quantiles (i.e., from Q50 to Q75). Thus, the
The efficiency of (total) assets (measured by net asset turnover)
results do not reject the research hypothesis, suggesting the existence
harms operating profitability for all quantiles, except Q10.
of a positive rela- tionship between climate change and firm
Furthermore, the negative effects are more pronounced for firms at the
profitability9.
lower (Q25) and higher (Q90) profitability quantiles than firms at the
The economic significance of the results shows that a 1℃ medium quantiles. Finally, the remaining firm-level control variable
degree (current ratio - CURR) does not have significant effects on firm
increase in temperature is related to a 1.30% increase in operating
profitability in any case. This result is in line with previous findings on
profitability (for the most profitable firms – Q90). However, the
the insignificant impact of liquidity on firm profitability [43,44,32].
magnitude of this impact is lower for less profitable firms (Q10 and
The empirical results show that the characteristics of the energy
Q25). The impact of temperature increase on the firm profitability is
market are important determinants of firm profitability. Contrary to the
propagated through the sales channel. Higher temperature means
initial expectations and previous results [30], the impact of market
higher cooling demand, especially in the summer period.
concentration is significantly negative only for low (Q10 – Q25) and
With one exception (Q90), the impact of the amount of rainfall on
high quantiles (Q90). The effect of energy prices on firm profitability
firm profitability is found to be statistically insignificant, a finding
is negative and statistically significant for all quantiles, suggesting that
which is consistent with previous literature on climate change [3].
regulated end-user prices for electricity and gas harm firm
Regarding firm characteristics, the results in Table 6 confirm those
performance. A similar negative relationship has been reported by
Jarait´e and Kaˇzukauskas (2013), but their results are statistically
insignificant. The
8
Specifically, I define five quantiles, namely Q10, Q25, Q50, Q75, and Q90.
9
In an unreported analysis, a non-linear relationship between temperature
and firm profitability has been tested. The empirical results do not confirm this kind of relationship.

11
S.G. Anton Applied Energy 295 (2021) 117051

Table 12
10
See Ku and Yen [33] for an extensive discussion on the relationship
between degreee of indebtedness and firm profitability.

12
 S.G. Anton
Table 8
Robustness checks.
A. Dependent variable: EBWA B. Dependent variable: ROE

OLS Q10 Q25 Q50 Q75 Q90 OLS Q10 Q25 Q50 Q75 Q90

TEMP 0.5503 0.9787** 0.5075*** 0.3115* 0.3359 1.3041*** 0.2036 0.6387 0.5363** 0.4627*** 0.3672** 0.6066**
(0.3692) (0.4650) (0.1579) (0.1805) (0.3724) (0.2606) (0.3887) (0.6376) (0.2483) (0.1388) (0.1567) (0.2754)
RAIN 0.0268 0.0310 0.0189 0.0343 —0.0258 —0.0719 —0.0468 —0.0983 0.0724** 0.0430** 0.0259 0.0441
(0.0473) (0.0471) (0.0212) (0.0227) (0.0390) (0.0471) (0.0715) (0.1234) (0.0286) (0.0207) (0.0269) (0.0579)
CFTA 56.920*** 79.387*** 75.991*** 68.841*** 50.429*** 42.398** 156.405*** 207.760*** 205.741*** 172.181*** 151.947*** 111.854***
(10.6728) (22.9701) (12.0418) (13.0437) (15.4318) (17.5656) (35.3763) (19.6107) (16.0470) (18.2253) (19.8380) (29.4942)
TQ 2.9834* —0.6393 1.9891 3.3773* 7.4104*** 6.3019** 2.4372 3.4127 3.9401** 4.1335*** 7.8420*** 14.8774***
(1.7606) (3.3070) (1.6401) (1.8012) (2.3229) (2.8513) (3.4361) (6.6787) (1.9408) (1.6001) (2.2733) (5.4386)
LEV 17.6962*** 2.8147 6.3938 19.6891*** 40.5778*** 40.3932*** 5.4168 —10.5505 5.1782 13.6370*** 27.8132*** 41.7643***
(6.4569) (7.2909) (4.1133) (3.8062) (4.7136) (4.7802) (9.6084) (8.9456) (4.5678) (3.2444) (3.4731) (5.0024)
SIZE1 0.4753 1.4354*** 0.6359*** —0.2232 —0.5546 —0.2486 1.1832* 2.1808*** 0.9925*** 0.4762*** —0.2397 —1.3868***
(0.3885) (0.4349) (0.2102) (0.2354) (0.3563) (0.7125) (0.6971) (0.7226) (0.2774) (0.1780) (0.2585) (0.3922)
8

NAT —1.8483*** —1.2783 —2.4761** —1.9195** —1.5185** —1.6389*** 0.4717 —6.0737** —1.5946 1.1323 2.7026*** 6.4047***
(0.6764) (1.5206) (0.9799) (0.8234) (0.7187) (0.4986) (0.9173) (3.0551) (1.1945) (0.8746) (0.8719) (1.9883)
CURR 0.0081 —0.0525 —0.1712 0.0301 0.3833 1.7496* 0.2275 0.0626 0.0275 0.2476* 0.2569* 0.0738
(0.3394) (0.9533) (0.2891) (0.3609) (0.4782) (0.9831) (0.2278) (0.2096) (0.1749) (0.1299) (0.1431) (0.1104)
MC —0.0887* —0.1495** —0.0477 —0.0175 —0.0115 —0.0671 —0.0319 —0.0093 0.0122 0.0038 0.0037 0.0130
(0.0532) (0.0669) (0.0316) (0.0204) (0.0329) (0.0602) (0.0715) (0.0770) (0.0231) (0.0145) (0.0207) (0.0344)
EP —39.2324*** —27.3952* —17.8117*** —31.7805*** —59.7183*** —91.8668*** —18.8360 —24.2976 9.0930 6.4856 12.2142 21.6542
(14.2444) (13.9823) (5.5159) (5.5989) (10.9578) (12.3141) (22.5249) (51.8214) (20.1770) (7.8627) (10.0491) (25.3299)
Constant —4.0921 —27.6120*** —10.7797*** 0.5341 9.9092 13.0191 —18.7875 —39.8109*** —37.5968*** —25.8065*** —16.1981*** —6.8848
(5.8561) (7.5798) (3.5743) (3.5871) (7.2121) (14.0653) (12.2930) (13.7696) (6.5197) (3.9766) (3.9015) (6.9419)
R2/Pseudo R2 0.2808 0.2125 0.1855 0.1691 0.1998 0.2132 0.3449 0.3565 0.3129 0.2705 0.2412 0.2456
Observations 763 763 763 763 763 763 763 763 763 763 763 763

Notes: The dependent variables are EBWA (panel A) and ROE (panel B). ***, ** and * stand for significance levels 1%, 5% and * at 10%. The numbers in parentheses are robust SEs for the OLS estimation and BSEs (QR
estimations). EBMA stands for operating profit margin; ROE - return on equity; TEMP – temperature; RAIN – rainfall; CFTA - cash flows; TQ - Tobin’s Q; SIZE1 - size measured as log of enterprise value; LEV -
leverage; CURR - current ratio; NAR - net asset turnover; EP - electricity price; MC - market concentration.

Applied Energy 295 (2021) 117051

S.G. Anton Applied Energy 295 (2021) 117051

significantly negative effects on the upper quantiles (Q75-Q90) are by the choice of firm profitability measure.
stronger than the significant negative effects on the lower quantiles
Overall, this study shows that not only firm-specific factors but also
(Q10-Q25).
energy market-related factors have a strong influence on the profit-
Overall, the empirical results confirm previous findings regarding
ability of listed firms from the energy and gas sectors. The findings of
the role of the amount of rainfall [3], firm size [30], or liquidityin
this study are particularly helpful for portfolio risk managers, energy
explaining differences in firm profitability. Unlike other studies, the
traders, and policymakers. For example, before designing industrial
current empirical analysis finds other key determinants for this partic-
policy initiatives in the energy and gas sectors, policy-makers should
ular sample of EU firms (market concentration and energy prices) or
consider the intended effects of policy intervention for their targets.
other signs for leverage (a positive impact).
Like any other study, this too has some limitations. Inferences
drawn from this study cannot be generalized to other countries from
4.3. Robustness checks different regions. Thus, the analysis can be extended to examine how
increased temperature affects firm profitability in different climatic
4.3.1. Alternative measure of profitability conditions such as Asian countries or South-American ones. Also, further
In my earlier analysis, I employ the operating profit margin research is needed to analyze the impact of climate change using
(EBMA) as the dependent variable. As a robustness check of my quarterly or monthly data. Furthermore, the time period of the analysis
empirical find- ings, I use in this sub-section the return on equity should be extended to incorporate the new evolution of the
(ROE) to capture the impact of rising global temperature on the temperature.
profitability expected by shareholders. The regression results from this
estimation are reported in Table 7. Declaration of Competing Interest
The results agree with the main estimations (Table 6) and provide
additional support for the non-homogeneous relationship between The authors declare that they have no known competing financial
increased temperature and firm profitability in the energy and gas sec- interests or personal relationships that could have appeared to
tors. However, we notice that the impact of temperature on return on influence the work reported in this paper.
equity is more important for the firms situated in the middle-range
quantiles. References
Also, the sign and significance of the controls are in line with the
baseline regression. Listed firms with higher growth opportunities and [1] Intergovernmental Panel on Climate Change. Global Warming of 1.5 ◦ C.
Switzerland, Geneva: 2018 October; 2018.
more cash flows exhibit higher profitability. With one exception
[2] Hugon A, Law K. Impact of climate change on firm earnings: evidence from and
(Q90), larger firms are more profitable. temperature anomalies. SSRN working paper; 2019.
[3] Addoum J, Ng D, Ortiz-Bobea A. Temperature shocks and establishment sales. Rev
4.3.2. Alternative measure of firm size Financ Stud 2020;33(3):1331–66. https://doi.org/10.1093/rfs/hhz126.
[4] International Monetary Fund. World Economic Outlook: A Long and Difficult
According to Dang et al. [45], firm size is found an important Ascent. Washington; 2020.
source of heterogeneity in firm profitability. Alternatively, the natural [5] Brabazon T, Idowu SO. Weather derivatives. Accountancy Ireland 7e9. Weather
loga- rithm of market capitalization is employed as an alternative Derivatives: A New Class of Financial Instruments; 2002.
[6] Nikolaou I, Evangelinos K, Leal Filho W. A system dynamic approach for
measure for firm size (SIZE1) in order to check the robustness of the exploring the effects of climate change risks on firms’ economic performance. J
main findings. The results of the robustness checks presented in Table Cleaner Prod 2015;103:499–506. https://doi.org/10.1016/j.jclepro.2014.09.086.
8 (panels A and B) confirm that the positive relationship between [7] Wieczorek-Kosmala M. Weather risk management in energy sector: The polish
case. Energies 2020;13:945. https://doi.org/10.3390/en13040945.
increased temperature and firm profitability remains unaltered when [8] Gallup JL, Sachs JD, Mellinger AD. Geography and economic development. Int
alternative measures of firm size are used in the analysis. Reg Sci Rev 1999;22:179–232.
Overall, the empirical results show that increased temperature has a [9] Nordhaus W. Geography and macroeconomics: new data and new findings. Proc
Natl Acad Sci 2006;103:3510–7.
significantly larger positive impact on high-performing firms than in
[10] Jones BF, Olken BA. Climate shocks and exports. Am Econ Rev 2010;100:454–9.
low-performing firms. Thus, the temperature effect is not [11] Hsiang SM. Temperatures and cyclones strongly associated with economic
homogeneous as reported in previous research on corporate finance production in the Caribbean and Central America. Proc Natl Acad Sci 2010;107:
[3]. Thus, findings from this sensitivity analysis corroborate our main 15367–72.
[12] Bansal R, Ochoa M. Temperature, aggregate risk, and expected returns. NBER
regression results reported in Table 6. Working Paper Series no. 17575; 2011.
[13] Dell M, Jones BF, Olken BA. Temperature shocks and economic growth: evidence
5. Conclusions from the last half century. Am Econ J: Macroecon 2012;4:66–95.
[14] Ghalehsard SK, Shahraki J, Akbari A, Shahraki AS. Assessment of the impacts of
climate change and variability on water resources and use, food security, and
The impact of climate change has received considerable attention economic welfare in Iran. Environ Dev Sustain 2021. https://doi.org/10.1007/
in the last decade. However, little is known about the impact of climate s10668-021-01263-w.
[15] Bansal R, Kiku D, Ochoa M. Climate change and growth risks. Working Paper,
change on firm profitability. The study fills this gap in the extant liter- Duke University.
ature providing empirical evidence on the impact of climate change on [16] Hong HG, Li FW, Xu J. Climate risks and market efficiency. J Economet 2017;208:
the financial performance in the energy and gas sectors using a sample 265–81.
[17] Jawid A. A Ricardian analysis of the economic impact of climate change on
of 147 listed firms active in 21 European Union countries. A panel
agriculture: evidence from the farms in the central highlands of Afghanistan.
quantile regression is employed in the paper in order to analyze the J Asian Econ 2020;67:101177. https://doi.org/10.1016/j.asieco.2020.101177.
impact at different quantiles of the distribution of firm profitability. [18] Cortignani R, Dell’Unto D, Dono G. Paths of adaptation to climate change in major
Italian agricultural areas: effectiveness and limits in supporting the profitability of
The empirical results suggest that the impact of increased annual
farms. Agric Water Manage 2021;244:106433. https://doi.org/10.1016/j.
temperature on firm profitability is not homogeneous among various agwat.2020.106433.
quantile levels: the increased temperature enhances the operating [19] Balvers R, Du D, Zhao X. Temperature shocks and the cost of equity capital:
margin profitability of low and high profitable firms. The relationship implications for climate change perceptions. J Bank Finance 2017;77:18–34.
https://doi.org/10.1016/j.jbankfin.2016.12.013.
between temperature and firm profitability is trivial when the mid- [20] Tzouvanas P, Kizys R, Chatziantoniou I, Sagitova R. Can variations in temperature
range performance quantiles are considered. As a robustness check, I explain the systemic risk of European Firms? Environ Resour Econonomics
use another measure for firm profitability (return on equity) and firm 2019; 74:1723–59. https://doi.org/10.1007/s10640-019-00385-0.
[21] Gernaat DEHJ, de Boer HS, Daioglou V, Yalew SG, Muller C, van Vuuren DP.
size (market capitalization). The estimates of the coefficients of Climate change impacts on renewable energy supply. Nat Clim Change 2021;11:
interest are qualitatively the same as for the main firm profitability 119–25. https://doi.org/10.1038/s41558-020-00949-9.
measure, only their magnitudes are slightly different. Hence, the
results are not driven

9
S.G. Anton Applied Energy 295 (2021) 117051

[22] van Ruijven BJ, De Cian E, Sue Wing I. Amplification of future energy
[34] Dell M, Jones BF, Olken BA. What do we learn from the weather? The new
demand growth due to climate change. Nat Commun 2019;10:2762.
climate- economy literature. J Econ Lit 2014;52:740–98.
https://doi.org/ 10.1038/s41467-019-10399-3.
[35] Kizhakethalackal ET, Mukherjee D, Alvi E. Quantile regression analysis of health-
[23] Henley A, Peirson J. Non-linearities in electricity demand and temperature:
aid and infant mortality: a note. Appl Econ Lett 2013;20:1197–201.
parametric versus non parametric methods. Oxford Bull Econ Stat
[36] Coad A, Rao R. Innovation and firm growth in high-tech sectors: A quantile
1997;59(1): 149–62.
regression approach. Res Policy 2008;37(4):633–48.
[24] Dischel, B., (2002), Introduction to the Weather Market: Dawn to Mid-Morning,
[37] Conyon MJ, He L. Firm performance and boardroom gender diversity: a quantile
[w:] Dischel B. (red.). Climate Risk and the Weather Market, Risk Books 2002, s.4.
regression approach. J Bus Res 2017;79:198–211. https://doi.org/10.1016/j.
[25] Weagley D. Financial sector stress and risk sharing: evidence from the weather
jbusres.2017.02.006.
derivatives market. Rev Financ Stud 2018;32(6):2456–97. https://doi.org/
[38] Koenker R, Bassett G. Regression quantiles. Econometrica 1978;46:33–50. https://
10.1093/rfs/hhy098.
doi.org/10.2307/1913643.
[26] Orbis. Orbis. Bureau van Dijk; 2019. Retrieved from https://orbis2.bvdep.com/.
[39] Maça˜s Nunes P, Neves Sequeira T, Serrasqueiro Z. Firms’ leverage and labour
[27] World Bank (2019). Climate Change Knowledge Portal, retrieved from
productivity: a quantile approach in Portuguese firms. Appl Econ 2007;39(14):
https:// climateknowledgeportal.worldbank.org/download-data.
1783–8. https://doi.org/10.1080/00036840600707076.
[28] Eurostat; 2019. Electricity market indicators, retrieved from https://ec.europa.eu/
[40] Buchinsky M. Estimating the asymptotic covariance matrix for quantile regression
eurostat/statistics-explained/index.php/Electricity_market_indicators.
models a Monte Carlo study. J Econ 1995;68:303–38.
[29] Lee BS, Li M-YL. Diversification and risk-adjusted performance: a quantile
[41] Hao L, Naiman DQ. Quantile regression. London: Sage Publications Inc; 2007.
regression approach. J Bank Finance 2012;36(7):2157–73. https://doi.org/
[42] Hair JF, Black WC, Babin BJ, Anderson RE. Multivariate Data Analysis. 7th ed.
10.1016/j.jbankfin.2012.03.020.
New York: Pearson; 2010.
[30] Jaraite˙ J, Kaˇzukauskas A. The profitability of electricity generating firms and
[43] Majumdar SK. The impact of size and age on firm-level performance: some
policies promoting renewable energy. Energy Econ 2013;40:858–65. https://doi.
evidence from India. Rev Ind Organ 1997;12:231–41.
org/10.1016/j.eneco.2013.10.001.
[44] Sur D, Chakraborty K. Evaluating relationship of working capital and profitability:
[31] Li M-YL, Yang T-H, Yu S-E. CEO stock-based incentive compensation and firm
a study of select multinational companies in the Indian pharmaceutical sector. IUP
performance: a quantile regression approach. J Int Financ Manage Account 2015;
J Manage Res 2011;10:7–22.
26:39–71. https://doi.org/10.1111/jifm.12022.
[45] Dang C, Li Z, Yang C. Measuring firm size in empirical corporate finance. J Bank
[32] Pervan M, Pervan I, C´urak M. Determinants of firm profitability in the Croatian
Finance 2018;86:159–76. https://doi.org/10.1016/j.jbankfin.2017.09.006.
manufacturing industry: evidence from dynamic panel analysis. Econ Res-
[46] Cronin J, Anandarajah G, Dessens O. Climate change impacts on the energy
Ekonomska Istraˇzivanja 2019;32(1):968–81. https://doi.org/10.1080/
system: a review of trends and gaps. Clim Change 2018;151:79–93.
1331677X.2019.1583587.
https://doi.org/ 10.1007/s10584-018-2265-4.
[33] Ku YY, Yen TY. Heterogeneous Effect of Financial Leverage on Corporate
[47] Hillier D, Pindado J, Queiroz Vd, Torre Cdl. The impact of country-level corporate
Performance: A Quantile Regression Analysis of Taiwanese Companies. Review of
governance on research and development. J Int Bus Stud 42(1):76–98.
Pacific Basin Financial Markets and Policies 19(3);2016: https://doi.org/10.1142/
S0219091516500156.

10