Economics 4312 Business Cycles and Forecasting

Topics for Review from Chapters 5-7 (Exam 2)

1. What is meant by model specification? Explain the role of economic theory. How is a mathematical specification determined that is used to estimate the assumed cause and effect relationship?

2. What is the role of a null hypothesis versus alternative hypothesis is testing your economic theory? Why is the only the slope coefficient important in a simple regression model in testing your assumed theoretical relationship?

3. Explain what is meant by the standard error of the regression coefficient. How is it different from the standard error of the estimate (regression)?

4. How are t-values used to verify the importance of a predictor variable in a regression model?

5. What is the relationship between the standard error of the estimate (regression) and the standard error of the forecast for small samples?

6. In a simple regression model how is a point forecast and an interval forecast determined for a given value of the independent variable?

7. What criteria that are useful in selecting predictor variables (theoretical, correlation with dependent variable, independent from other predictor variables) can be illustrated with a correlation matrix?

8. How are partial regression coefficients interpreted in a linear multiple regression model? in a logarithmic multiple regression model?

9. Why is multicollinearity a problem in simulation but not in forecasting?

10. What is a beta coefficient and how can it be used to determine the relative importance of predictor variables?

11. How can a plot of residuals that are autocorrelated be used to identify specification error? Illustrate with error terms that suggest a need to correct for (a) trend influences, (b) cyclical influences, and (c) irregular influences.

12. What is meant by multicollinearity? How can it be detected? What is its effect of regression results? How can it be eliminated?

13. What is meant by positive first order autocorrelation (of error terms)? Why is the elimination of autocorrelation so important in forecasting? How is the Durbin-Watson statistic used to test for autocorrelation? How can autocorrelation be eliminated in regression models (give two methods)?

17. What is the usefulness of proxy variables?

18. How can dummy variables be used to correct for qualitative influences? for seasonal changes?

19. What is the difference between simple lags and distributed lags?