Consider again the post-1950 trend in U.S. GDP we examined in Exercise 61. Here are regression output and a residual plot when we use the log of GDP in the model. Is this a better model for GDP? Explain. Would you want to consider a different re-expression? If so, which one?

Transcribed Image Text:

Residuals 0.025 0.000 -0.025 -0.050- Dependent variable is: LogGDP R-squared = 99.0% s = 0.0268 Variable Intercept Years since 1950 0.50 0.75 Coefficient 0.3918 0.01358 Predicted 1.00 1.25