The Reservoir Water Level Data Analysis project aims to examine the impact of economic output and income inequality on the Reservoir Water Level. The following steps were taken to complete the enhanced version of this project:
The original regression model was enhanced by incorporating indicators for power inequality, including literacy rate, mortality rate, per capita income, unemployment, and rainfall. These variables were added to the model to better explain the variation in the environmental quality indicator.
Instead of using district-level data for the Gini index, state-census-year level data provided by Pandey and Gautam (2020) was utilized to test how this substitution of variables alters the results.
The role of standard errors in regression analysis was clearly articulated for the context of this study. Standard errors were utilized to measure the precision of the estimated coefficients and to test hypotheses about population parameters.
A Chow Test and a t-test were conducted to statistically test for a structural break in the mean environmental quality across different state-groups. The inference drawn from these tests was compared.
The enhanced regression model was used to run a Monte Carlo simulation procedure to comment on the consistency of OLS estimates. The intercept and slope coefficient estimates obtained were treated as true population parameters.
The maximum likelihood strategy was considered as a means to improve the estimation of the enhanced regression model.
It was found that the average environmental quality differs significantly across state groups. A test was conducted to examine the variance in environmental quality across the state-groups. If the variance differs significantly across state groups, the assumption of OLS is violated. An alternative estimation strategy was suggested to account for such a variance structure.