Question

We are estimating a Cobb. Deras production function. The variables are as follows Suppose we are estimating a Cobb-Doug production production of wheat in metric tons land land area in acres labor number of labor capital number of tractors You run the log-log model, In production) constant + a Inland) +hInilabor) + c.In(capital) + and get the following results production Coell. land) 0.27 In labor) 0.32 la capital) 0.40 constant 2.03 R = 0.92 1.56 2.45 3.33 1.02 P> 0.15 0.03 001 033 95% conf. Interval 0.18 0.35 0.22 0.38 0.36 0.42 1.75 P-value (F) = 0.0000 12.48 8. What does the coefficient of Inland) indicate? Write down in plain English. 9. Is land a significant factor to explain production? Explain. Will you exclude land from the model? Why or why not? 10. The R of the second model (0.92) is higher than the first model (0.56). Can you say that the second model is better than the first model? Explain.

Answer 1

8) coefficient implies that the log of production is directly or positively related to the log of area in acres i.e. higher the area, higher will be the production.

9) land is not a significant factor in explaining production as the p-value of the coefficient is .15. For the variable to be significant, the p-value must be below the significance level of 10%(0.1), 5%(0.05) or 1%(0.05) as desired. As the variable is not significant, it can be excluded from the model as it does represent represent significant relationship.

10) Higher R square generally implies a better model, however, it's not necessarily so as the higher value can be due to over-fitting of the data and variables and the resulting model might not have any basis or underlying logic.