What would you do if you had to choose between a model that satisfies all statistical criteria but does not satisfy economic theory and a model that fits established economic theory but does not fit many statistical criteria? Briefly explain your answer.
In measuring returns to scale in electricity supply, cross-sectional data of 145 privately owned utilities in the United States is used to regress the log of total cost on the logs of output, wage rate, price of capital, and price of fuel. Suppose the residuals estimated from this regression exhibit "serial" correlation, as judged by the Durbin–Watson d. The figure below plots the estimated residuals on the log of output.
What does the figure show?
How can you get rid of "serial" correlation in the preceding situation?
