Applied economic - Assignment Example

Given that model 2 is the correct model, the Gauss-Markov theorem tells me that the model’s ordinary least square estimators are the best estimators under classical linear regression models. This implies that the coefficients are the best for the model and the R2 is the highest that the model can attain (Wooldridge 102). Suppose a test was conducted on the hypothesis at 15 percent level of confidence and a p-value of 0.119 obtained, the test result would mean that no significant relationship exist.

This is because for a two tailed test, implied by the alternative hypothesis that β1≠-3, then the confidence level for the test is 0.075. The p-value is therefore greater that the confidence level and this means that the null hypothesis is not rejected. Conducting the test 10 percent level could change the conclusion on significance. This is because of the associated increment in confidence interval that could extend to cover the critical value. Assuming that Ram (2009) is correct in his model in which openness is a significant factor to government’s expenditure, then expected value of error term in the model that omitted the openness variable cannot be zero and the expression is not valid.

This is because of effects openness that was factored in the error term and meant that what was perceived to be the error term was not error but effects of an unidentified variable. Even though both models leads to rejection of the null hypothesis, the greater absolute t-statistic from model two offers better evidence for rejecting the null hypothesis because it would ensure rejection of the null hypothesis at a greater level of precision than would model 1. The greater absolute value of the statistic in model two explains this because greater statistic than a reference value implies significance.

Suppose population is added to model 1, then