In order to have full access of this Article, please email us on thedocumentco@hotmail.co.uk
Data Analysis
The following model is taken into consideration for the data analysis of this research:
Y= 𝛼0 + 𝛼1Tt + 𝛼2EG𝑡 +a3EC𝑡 + 𝛼4AE𝑡 + 𝜇𝑡 (1)
After collecting a data set from 1990 to 2001 on all the above mentioned variables, the coefficients were estimated by running a regression in SPSS and deriving the entire econometrics for analysis to be more meaningful. The following is the estimated model:
Y = 6.314 + 0.184Tt + 0.087EG𝑡 +0.209EC𝑡 + 0.484AE𝑡 + 𝜇𝑡
Interpretation of coefficients:
6.314: When all the other factors such as time, electricity generation, electricity consumption and access to electricity is kept 0 then electricity productivity will increase by 6.314%.
0.184: Time is the dummy variable and it takes two values 0 and 1. When the year taken is after reforms T takes the value of 1 and hence shows an increase in electricity performance by 0.184% keeping electricity generation, electricity consumption and access to electricity constant. And if the year is before reforms T takes the value 0 showing no effect on electricity productivity.
0.087: 1 billion kWhr increase in electricity generation will bring about 0.087% increase in electricity productivity, keeping other variables such as electricity consumption and access to electricity constant.
0.484: When there is a 1 unit increase in the access to electricity the electricity productivity increases by 0.484% keeping other variables like electricity generation and electricity consumption constant.
Besides this the Pearson correlation test suggests the statistical evidence of the linear relationship among the same pairs of variables in the population. In this case all the pearson’s values are positive and shows the variables are positively correlated to each other. Whereas r-values being closer to 1 shows the two variables are strongly correlated to each other and that one affects the other whereas r-value being closer to 0 shows that they are weakly correlated. According to the results of econometrics,
Table 1
Y is strongly correlated to EG, EC and AE; and weakly correlated to T.
T is weakly correlated Y, EG, EC & AE.
EG is strongly correlated to Y, EC & AE whereas weakly correlated to T.
EC is strongly correlated to Y, EG & AE whereas weakly correlated to T.
AE is strongly correlated to Y, EG & EC whereas weakly correlated to T.
In order to check the significance of the model, the fitness of good would be verified through different parameters such as indicated in table 2:
Table 2
The model has R-square = 0.912 and Adjusted R-square = 0.91239 indicating that approximately 91% of the variation in the model is explained by the variables i.e. T, EG, EC & AE taken into account and the remaining 9% is unexplained i.e. included as error. The F-statistic is highly significant thus the model explains the significant amount of variance in the electricity performance over the past years in Argentina.
Consequently, autocorrelation of residuals over different time intervals is very common in time series. Hence Durbin Watson test checks the presence of autocorrelation in any model. In this case the Durbin Watson value is 1.221; showing the existence of positive autocorrelation in this model.
Analyzing model 2:
As discussed in chapter 3 about the following model, when estimated gave the different values for the coefficients:
Y= 𝛼0 + 𝛼1Tt + 𝛼2EG𝑡 + a3EC𝑡 +a4AE𝑡 + 𝛼5(T𝑡)(EGt)+ 𝛼6(T𝑡)(ECt) +𝛼7(T𝑡)(AEt) + 𝜇𝑡 (2)
Y = 6.314 + 0.184T + 0.087EG + 0.209EC + 0.484AE + 0.314(T)(EG) + 0.319(T)(EC)+ 0.186(T)(AE)
The term T*EG shows the multiplicative interaction of the dummy variable T with EG and the coefficient with it is called the average treatment effect. Hence 0.314 shows the net effect of reforms and electricity generation had on the productivity and performance of electricity in Argentina. Similarly, 0.319 shows the net effect of reforms and electricity consumption whereas 0.186 shows the net effect of reforms and access to electricity.
Carrying on the Pearson test, resulted in the displaying the strong and weak correlation between the two variable with being perfectly correlated to itself. As the r-value approaching towards 1 shows the strong correlation and r-value approaching towards…
Recent Comments