Page 18 - ECLECTIC MARCH-2022 INSIDE PAGES_p2.indd
P. 18
regression model. Lower significance levels indicate Table 2: Coefficients P-value
that stronger evidence before is required to reject the null 0.0000 1.0000
hypothesis. For the purpose of the study the significance Intercept 0.3372 0.0056 *
level is 5% or 0.05. Import –0.5315 0.0005 *
Despatch –0.4052 0.0063 *
However, it is worthwhile to note that there are Demand –0.0015 0.9889
certain limitations to this study. The article essentially OD
attempts to study the movement of the variables
during a period of shock. Hence, by nature, such a Regression Statistics
study will have restrictions on the size of the sample
set. This limits the accuracy of the model in terms of R Square 0.85747
the coefficients of the model (βi). However, the models
being considered here is solely for the purpose of Adjusted R Square 0.81675
identifying the movement and dependency of the
dependent variable on the independent variables, that F-stat 21.05707
is, the sign (+/–) of the βi’s and the p-value of the null
hypothesis that the βi = 0. Significance F 8.37E-06
Observations 19
* indicates that the respective variables are significant at a
significance level of 0.05
RESULT As seen from the regression statistics, Import, Despatch
and Demand are the variables that affect the movement
Prior to the multiple linear regression model, we test for of Stock of coal in days at an aggregate level across India,
correlation between the independent variables. Table 1 as evidenced by the p-value. The F-statistic confirms
shows the correlation matrix between the independent that the regression model is improved by the inclusion
variables. There is a moderate and positive correlation of the independent variables. We note that the variable
between Demand and Despatch, and similarly for OD corresponding to overdues do not affect the aggregate
and Demand. The correlation between Despatch and coal stock position, as evidenced by the corresponding
Demand is intuitive, as it shows that with increasing p-value. This confirms the hypothesis that increasing
power demand, there is a pull for coal from suppliers, overdues to gencos (and subsequently to coal suppliers)
thereby increasing the despatch volumes. However, we did not affect the coal stock of power plants during this
see that for none of the independent variables there is a period. This result is expected, as increasing overdues
strong correlation with the others. is unlikely to lead to curtailment of supplies from coal
companies.
Table 1: Import Despatch Demand OD The results show that the coefficient of the variable
1 1 corresponding to demand is negative signifying that with
Import 1 1 increased demand, there is depletion in the coal stock.
Despatch – 0.0750 0.4968 0.3561 This result is as expected, as an increase in power demand
Demand – 0.2027 0.1005 is likely to be met by higher coal consumption. This trend
OD – 0.0931 can also be observed graphically as shown in Figure 1.
Similarly, the coefficient of the variable corresponding
We proceed to formulate the multiple linear regression to volume of imported coal is positive, showing that a
model with these independent variables. The summary decrease in the volume of imported coal is associated
statistics of the multiple regression model is laid out in with a decrease in number of days of coal stock with
Table 2. thermal power plants.
18
that stronger evidence before is required to reject the null 0.0000 1.0000
hypothesis. For the purpose of the study the significance Intercept 0.3372 0.0056 *
level is 5% or 0.05. Import –0.5315 0.0005 *
Despatch –0.4052 0.0063 *
However, it is worthwhile to note that there are Demand –0.0015 0.9889
certain limitations to this study. The article essentially OD
attempts to study the movement of the variables
during a period of shock. Hence, by nature, such a Regression Statistics
study will have restrictions on the size of the sample
set. This limits the accuracy of the model in terms of R Square 0.85747
the coefficients of the model (βi). However, the models
being considered here is solely for the purpose of Adjusted R Square 0.81675
identifying the movement and dependency of the
dependent variable on the independent variables, that F-stat 21.05707
is, the sign (+/–) of the βi’s and the p-value of the null
hypothesis that the βi = 0. Significance F 8.37E-06
Observations 19
* indicates that the respective variables are significant at a
significance level of 0.05
RESULT As seen from the regression statistics, Import, Despatch
and Demand are the variables that affect the movement
Prior to the multiple linear regression model, we test for of Stock of coal in days at an aggregate level across India,
correlation between the independent variables. Table 1 as evidenced by the p-value. The F-statistic confirms
shows the correlation matrix between the independent that the regression model is improved by the inclusion
variables. There is a moderate and positive correlation of the independent variables. We note that the variable
between Demand and Despatch, and similarly for OD corresponding to overdues do not affect the aggregate
and Demand. The correlation between Despatch and coal stock position, as evidenced by the corresponding
Demand is intuitive, as it shows that with increasing p-value. This confirms the hypothesis that increasing
power demand, there is a pull for coal from suppliers, overdues to gencos (and subsequently to coal suppliers)
thereby increasing the despatch volumes. However, we did not affect the coal stock of power plants during this
see that for none of the independent variables there is a period. This result is expected, as increasing overdues
strong correlation with the others. is unlikely to lead to curtailment of supplies from coal
companies.
Table 1: Import Despatch Demand OD The results show that the coefficient of the variable
1 1 corresponding to demand is negative signifying that with
Import 1 1 increased demand, there is depletion in the coal stock.
Despatch – 0.0750 0.4968 0.3561 This result is as expected, as an increase in power demand
Demand – 0.2027 0.1005 is likely to be met by higher coal consumption. This trend
OD – 0.0931 can also be observed graphically as shown in Figure 1.
Similarly, the coefficient of the variable corresponding
We proceed to formulate the multiple linear regression to volume of imported coal is positive, showing that a
model with these independent variables. The summary decrease in the volume of imported coal is associated
statistics of the multiple regression model is laid out in with a decrease in number of days of coal stock with
Table 2. thermal power plants.
18
