278 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
Models for Hypothesis Testing
To examine the relationship between sustainability behav-
iors aimed at carbon emission reduction and firm financial
performance, as described in the hypothesis, a dynamic
panel data model was employed.
.YEAR
CFP EE SIZE
LEVARAGE
i
Behavior
.
,,t ,t
,t
,t
t
f
i f,t
3 5
6
11
a b b
b b8COMMODITY
b
+/b10Sustainability
=++
++
+(1)
CFP ROA or Tobin Q
,,t ,t t =(2)
where
CFPf,t =denotes the corporate financial performance for
company f at time t being measured by CFP
represented by ROA (model 1) and Tobin’s Q
(model 2)
EEf,t–1 =Energy reduction for company f at time t
SIZEf,t =Natural log of assets of each firm for company f
at time t
LEVARAGEf,t =Level of risk of the company’s operation
for company f at time t
COMMODITYf =Dummy variable of mining commodity
sector the company belongs to 0 otherwise.
i.YEARf Yearly, dummy, variables for com-
pany f.
Data and Sample Collection
This study employed secondary data analysis to explore
the relationship between sustainability behavior, specifi-
cally carbon emission reduction initiatives, and financial
performance in the mining industry. The GOLDEN
Sustainability Database (version 3.0) provided data on sus-
tainability initiatives across two decades for approximately
9,000 publicly traded companies, including mining firms
(Cenci et al., 2023b, 2023a). This data aligns with the 17
Sustainable Development Goals and 14 sustainability activ-
ities, with particular emphasis on emission reduction initia-
tives (Cenci et al., 2023b, 2023a).. Financial performance
and mining commodity information for each company
were concurrently retrieved from the Bloomberg terminal,
renowned for its historical financial data for global com-
panies dating back to 1980 (Cenci et al., 2023b, 2023a).
Sustainability behavior and historical financial data were
collected for mining companies across diverse commodi-
ties to ensure sample representativeness. Data from both
sources were combined to create an unbalanced panel
dataset of companies spanning 2010–2020, with the final
sample size determined by data availability and statistical
considerations.
Variables
Dependent Variables
Financial performance (FP) was measured using two estab-
lished indicators: Tobin’s Q and ROA. Tobin’s Q reflects
market expectations, capturing intangible value through
market capitalization, book value, and ROA itself (Zahid
et al., 2022). A higher ratio suggests a positive market out-
look, influenced by factors like reputation, investor trust,
and risk (Zahid et al., 2022). ROA, on the other hand,
measures a company’s profitability and asset efficiency
(Zahid et al., 2022). A higher ROA indicates better utiliza-
tion of assets to generate profits.
Independent Variables
In this study, the independent variables are the diverse sus-
tainability behaviors adopted by firms and their energy effi-
ciency efforts. Sustainability behaviors are categorized based
on activity type (product development, donations, etc.)
and the targeted Sustainable Development Goal (SDG).
Table 2 details these variables. Energy efficiency, calculated
as shown in equation [3], is included as a separate variable.
EE
Sales
EC
Sales
EC
Sales
EC
,t
,t
,t
,t
,t
,t
,t
1
1
1
1
=
-
-
-
-
-e o
(3)
where EC represents total energy consumption in giga-
joules divided by net sales (Sales/Revenue) for company f
at time t or at time t–1. The variable EEf,t represents the
change in energy consumption for a specific company f at
the end of year t, adjusted by sales.
Control Variables
To ensure the robustness of the analysis and control for
potential confounding factors, this study incorporated sev-
eral common control variables, namely firm size, leverage,
and fixed effects (year and commodity). Firm size, measured
by the natural logarithm of total assets, was included due to
its established impact on financial performance. Leverage,
represented as total debt over total assets, was considered
for its potential influence on financial outcomes. The inclu-
sion of fixed effects (year and commodity) aimed to account
for unobserved factors, such as industry-specific trends or
economic fluctuations, and minimize their impact on the
relationship between carbon emission reduction practices
and financial performance.
Models for Hypothesis Testing
To examine the relationship between sustainability behav-
iors aimed at carbon emission reduction and firm financial
performance, as described in the hypothesis, a dynamic
panel data model was employed.
.YEAR
CFP EE SIZE
LEVARAGE
i
Behavior
.
,,t ,t
,t
,t
t
f
i f,t
3 5
6
11
a b b
b b8COMMODITY
b
+/b10Sustainability
=++
++
+(1)
CFP ROA or Tobin Q
,,t ,t t =(2)
where
CFPf,t =denotes the corporate financial performance for
company f at time t being measured by CFP
represented by ROA (model 1) and Tobin’s Q
(model 2)
EEf,t–1 =Energy reduction for company f at time t
SIZEf,t =Natural log of assets of each firm for company f
at time t
LEVARAGEf,t =Level of risk of the company’s operation
for company f at time t
COMMODITYf =Dummy variable of mining commodity
sector the company belongs to 0 otherwise.
i.YEARf Yearly, dummy, variables for com-
pany f.
Data and Sample Collection
This study employed secondary data analysis to explore
the relationship between sustainability behavior, specifi-
cally carbon emission reduction initiatives, and financial
performance in the mining industry. The GOLDEN
Sustainability Database (version 3.0) provided data on sus-
tainability initiatives across two decades for approximately
9,000 publicly traded companies, including mining firms
(Cenci et al., 2023b, 2023a). This data aligns with the 17
Sustainable Development Goals and 14 sustainability activ-
ities, with particular emphasis on emission reduction initia-
tives (Cenci et al., 2023b, 2023a).. Financial performance
and mining commodity information for each company
were concurrently retrieved from the Bloomberg terminal,
renowned for its historical financial data for global com-
panies dating back to 1980 (Cenci et al., 2023b, 2023a).
Sustainability behavior and historical financial data were
collected for mining companies across diverse commodi-
ties to ensure sample representativeness. Data from both
sources were combined to create an unbalanced panel
dataset of companies spanning 2010–2020, with the final
sample size determined by data availability and statistical
considerations.
Variables
Dependent Variables
Financial performance (FP) was measured using two estab-
lished indicators: Tobin’s Q and ROA. Tobin’s Q reflects
market expectations, capturing intangible value through
market capitalization, book value, and ROA itself (Zahid
et al., 2022). A higher ratio suggests a positive market out-
look, influenced by factors like reputation, investor trust,
and risk (Zahid et al., 2022). ROA, on the other hand,
measures a company’s profitability and asset efficiency
(Zahid et al., 2022). A higher ROA indicates better utiliza-
tion of assets to generate profits.
Independent Variables
In this study, the independent variables are the diverse sus-
tainability behaviors adopted by firms and their energy effi-
ciency efforts. Sustainability behaviors are categorized based
on activity type (product development, donations, etc.)
and the targeted Sustainable Development Goal (SDG).
Table 2 details these variables. Energy efficiency, calculated
as shown in equation [3], is included as a separate variable.
EE
Sales
EC
Sales
EC
Sales
EC
,t
,t
,t
,t
,t
,t
,t
1
1
1
1
=
-
-
-
-
-e o
(3)
where EC represents total energy consumption in giga-
joules divided by net sales (Sales/Revenue) for company f
at time t or at time t–1. The variable EEf,t represents the
change in energy consumption for a specific company f at
the end of year t, adjusted by sales.
Control Variables
To ensure the robustness of the analysis and control for
potential confounding factors, this study incorporated sev-
eral common control variables, namely firm size, leverage,
and fixed effects (year and commodity). Firm size, measured
by the natural logarithm of total assets, was included due to
its established impact on financial performance. Leverage,
represented as total debt over total assets, was considered
for its potential influence on financial outcomes. The inclu-
sion of fixed effects (year and commodity) aimed to account
for unobserved factors, such as industry-specific trends or
economic fluctuations, and minimize their impact on the
relationship between carbon emission reduction practices
and financial performance.