284 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
the consistency of findings, reinforcing the reliability of
observed associations between sustainability behaviours
and ROA.
Hypothesis 1a—Association between specific
sustainability behaviours and financial performance
(ROA):
• Null Hypothesis (H0): There is no statistically sig-
nificant association between specific sustainability
behaviours and higher financial performance (ROA).
• Alternative Hypothesis (H1): There is a statistically
significant association between specific sustainability
behaviours and higher financial performance (ROA).
ROA EE SIZE
COMMODITY i.YEAR
Sustainability Behavior
,,t ,t
,t
t
f f,t
i.f
1 2
4 5
a b b
b b
+/bi
=++
+b3LEVARAGE
++
where
ROAf,t =denotes the lagged corporate financial perfor-
mance for company f at time t being ROA
measure (model 1) and Tobin’s Q (model 2)
EEf,t =Energy reduction for company f at time t
SIZEf,t =Natural log of assets of each firm for company f
at time t
LAVARAGEf,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,t =Yearly, dummy, variables for company f.
Entity and Time Fixed Effects Regression
Introducing entity and time fixed effects in Model 1 (ROA)
significantly improved the model fit (R-squared: 0.479
vs. 0.426 without time effects). The F-test statistic also
increased, confirming a better data fit (Table 7). Control
variables “size” and “energy_efficiency” remained significant
(p =0.021, 0.017). Total Associations (p =0.028) maintained
Table 7. Regression results entity and time fixed effects model
return_on_asset Coef. St.Err. t-value p-value [95% Conf Interval] Sig
fncl_lvrg –5.04 2.26 –2.23 0.034 –9.663 –.417 **
size 6.24 2.555 2.44 0.021 1.014 11.466 **
energy_efficiency –6.385 2.529 –2.53 0.017 –11.557 –1.214 **
total_adoptofstand~s –0.33 4.139 –0.08 0.937 –8.795 8.135
total_associations –1.761 0.762 –2.31 0.028 –3.32 –0.202 **
total_communications 1.287 0.594 2.17 0.039 0.072 2.502 **
total_donationfund~s 0.242 0.651 0.37 0.713 –1.09 1.574
total_assetmodific~s 0.305 0.451 0.68 0.504 –0.618 1.228
total_incentives –2.418 1.273 –1.90 0.068 –5.021 0.186 *
total_modification~s 0.291 0.472 0.62 0.542 –0.674 1.256
total_newproducts –1.219 1.289 –0.95 0.352 –3.856 1.418
total_orgstructuring 0.159 1.825 0.09 0.931 –3.572 3.891
total_rdinvestments 0.561 1.144 0.49 0.627 –1.778 2.901
total_training 0.114 0.615 0.19 0.854 –1.143 1.372
total_volunteerism 2.644 3.614 0.73 0.47 –4.747 10.036
total_assess_measu~t –0.835 0.387 –2.16 .039 –1.628 –0.043 **
2020b 0
2021 8.127 3.2 2.54 .017 1.582 14.672 **
Constant –45.752 26.282 –1.74 .092 –99.504 8
Mean dependent var 5.141 SD dependent var 9.385
R-squared 0.479 Number of obs 97
F-test 125.170 Prob F 0.000
Akaike crit. (AIC) 560.278 Bayesian crit. (BIC) 604.048
***p.01, **p.05, *p.1
the consistency of findings, reinforcing the reliability of
observed associations between sustainability behaviours
and ROA.
Hypothesis 1a—Association between specific
sustainability behaviours and financial performance
(ROA):
• Null Hypothesis (H0): There is no statistically sig-
nificant association between specific sustainability
behaviours and higher financial performance (ROA).
• Alternative Hypothesis (H1): There is a statistically
significant association between specific sustainability
behaviours and higher financial performance (ROA).
ROA EE SIZE
COMMODITY i.YEAR
Sustainability Behavior
,,t ,t
,t
t
f f,t
i.f
1 2
4 5
a b b
b b
+/bi
=++
+b3LEVARAGE
++
where
ROAf,t =denotes the lagged corporate financial perfor-
mance for company f at time t being ROA
measure (model 1) and Tobin’s Q (model 2)
EEf,t =Energy reduction for company f at time t
SIZEf,t =Natural log of assets of each firm for company f
at time t
LAVARAGEf,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,t =Yearly, dummy, variables for company f.
Entity and Time Fixed Effects Regression
Introducing entity and time fixed effects in Model 1 (ROA)
significantly improved the model fit (R-squared: 0.479
vs. 0.426 without time effects). The F-test statistic also
increased, confirming a better data fit (Table 7). Control
variables “size” and “energy_efficiency” remained significant
(p =0.021, 0.017). Total Associations (p =0.028) maintained
Table 7. Regression results entity and time fixed effects model
return_on_asset Coef. St.Err. t-value p-value [95% Conf Interval] Sig
fncl_lvrg –5.04 2.26 –2.23 0.034 –9.663 –.417 **
size 6.24 2.555 2.44 0.021 1.014 11.466 **
energy_efficiency –6.385 2.529 –2.53 0.017 –11.557 –1.214 **
total_adoptofstand~s –0.33 4.139 –0.08 0.937 –8.795 8.135
total_associations –1.761 0.762 –2.31 0.028 –3.32 –0.202 **
total_communications 1.287 0.594 2.17 0.039 0.072 2.502 **
total_donationfund~s 0.242 0.651 0.37 0.713 –1.09 1.574
total_assetmodific~s 0.305 0.451 0.68 0.504 –0.618 1.228
total_incentives –2.418 1.273 –1.90 0.068 –5.021 0.186 *
total_modification~s 0.291 0.472 0.62 0.542 –0.674 1.256
total_newproducts –1.219 1.289 –0.95 0.352 –3.856 1.418
total_orgstructuring 0.159 1.825 0.09 0.931 –3.572 3.891
total_rdinvestments 0.561 1.144 0.49 0.627 –1.778 2.901
total_training 0.114 0.615 0.19 0.854 –1.143 1.372
total_volunteerism 2.644 3.614 0.73 0.47 –4.747 10.036
total_assess_measu~t –0.835 0.387 –2.16 .039 –1.628 –0.043 **
2020b 0
2021 8.127 3.2 2.54 .017 1.582 14.672 **
Constant –45.752 26.282 –1.74 .092 –99.504 8
Mean dependent var 5.141 SD dependent var 9.385
R-squared 0.479 Number of obs 97
F-test 125.170 Prob F 0.000
Akaike crit. (AIC) 560.278 Bayesian crit. (BIC) 604.048
***p.01, **p.05, *p.1