14.5 Exercises
WDI and Heteroscedasticity (***): Using the data in
wdi
, estimate the following equation for the year 2018 and report the results: fertrate=β0+β1⋅gdp+β2⋅litrate Conduct a test for heteroscedasticity of your choice. Is there heteroscedasticity present in the model? If yes, execute additional calculations to correct for it and report the results.Indy Homes III (***): The data
indyhomes
contains home values of two ZIP codes in Indianapolis. The model estimates the home value (dependent variable) based on a set of independent variables. The variables levels and garage refers to the number of stories and garage spots, respectively. The remaining variables are self-explanatory.- Create a dummy variable called northwest for the 46268 ZIP code.
- Report and interpret the results of the regression equation that uses ln(price) as the dependent variables and the folloing as independent variables: ln(sqft), northwest, ln(lot), bed, garage, and levels. In addtion, include an interaction between northwest and levels.
- What is the expected home value of a house in the 46228 ZIP code area with the following characteristics: 1900 sqft, 0.65 acres lot, 3 bedrooms, 3 bathrooms, 2 garage spots, and 2 story.
- Conduct a Breusch-Pagan-Godfrey test for heteroscedasticity. What do you conclude?
- Estimate the above model with heteroscedasticity-consistent (HC) standard errors. What changes compared to the model from Part b?
WDI and Multicollinearity (***): Use the command
subset()
on the WDI data and to select the variables fertrate, gdp, litrate, lifeexp, and mortrate for the year 2015. Estimate the following model fertrate=β0+β1⋅gdp+β2⋅litrate+β3⋅lifeexp+β4⋅mortrate Interpret the results. What do you conclude in terms of statistical significance and the value of R2? Use the functionvif
from the packagecar
. What can you say about the issue of multicollinearity in this case? Correct the issue of multicollinearity by adjusting your model.Wages, Productivity, and Autocorrelation (***):