Assignment:
1. Regress the variable "CO2" on the variable "population." Then, save the residuals by clicking Save Residuals in the output window. Now regress the residuals on the lagged residuals and a constant. You can do this by specifying the residuals as dependent variable, and clicking "lags" and "lags of dependent variables." Explain why the output of the regression suggests that the regression suffers front an autocorrelation problem.
2. Itegrms the variable "CO2" on the lag of the variable "CO2", the variable "population," and the lag of the variable "population" (i.e., include those three variables in your regression as In4nssors). Is the lag of "CO2" statistically significant at the 5% level? Is the lag of "population" statistically significant at the 5% level? Motivate your answers.
3. Using the regression of the previous question, predict CO2 emissions for 1986 (i.e. one period out of sample) under the assumption that population stays at 4.85539e+09, as it was in 1985. Note that CO2 emissions in 1985 was at 11238.68. Show your calculation.
4. Now regress the first difference of CO2 emissions on the first difference of population. What is the t-ratio in this regression?
5. For what purpose do we sometimes check the checkbox titled "robust" in the regression window of Cretl'? (Note: any answer that essentially repeats the word "robust" without adding anything essential will not receive points)
6. What is the hypothesis that is tested by the fitted for overall significance?
7. What is a disadvantage of the White test for hetcroskedasticity?