The document provides data on 81 cars including miles per gallon (MPG), horsepower (HP), weight (WT), top speed (SP), and cab space (VOL). A linear regression model is estimated with MPG as the dependent variable and SP, HP, and WT as independent variables. The model finds that MPG is positively related to HP and negatively related to SP and WT, with all coefficients being statistically significant. A White test is then conducted and finds evidence of heteroscedasticity in the error terms as would be expected with a cross-sectional data set involving diverse cars.
The document provides data on 81 cars including miles per gallon (MPG), horsepower (HP), weight (WT), top speed (SP), and cab space (VOL). A linear regression model is estimated with MPG as the dependent variable and SP, HP, and WT as independent variables. The model finds that MPG is positively related to HP and negatively related to SP and WT, with all coefficients being statistically significant. A White test is then conducted and finds evidence of heteroscedasticity in the error terms as would be expected with a cross-sectional data set involving diverse cars.
The document provides data on 81 cars including miles per gallon (MPG), horsepower (HP), weight (WT), top speed (SP), and cab space (VOL). A linear regression model is estimated with MPG as the dependent variable and SP, HP, and WT as independent variables. The model finds that MPG is positively related to HP and negatively related to SP and WT, with all coefficients being statistically significant. A White test is then conducted and finds evidence of heteroscedasticity in the error terms as would be expected with a cross-sectional data set involving diverse cars.
The document provides data on 81 cars including miles per gallon (MPG), horsepower (HP), weight (WT), top speed (SP), and cab space (VOL). A linear regression model is estimated with MPG as the dependent variable and SP, HP, and WT as independent variables. The model finds that MPG is positively related to HP and negatively related to SP and WT, with all coefficients being statistically significant. A White test is then conducted and finds evidence of heteroscedasticity in the error terms as would be expected with a cross-sectional data set involving diverse cars.
Download as DOCX, PDF, TXT or read online from Scribd
Download as docx, pdf, or txt
You are on page 1/ 7
Question No.
Table below gives data on 81 cars about MPG (average miles per gallons), HP (engine horsepower), VOL (cubic feet of cab space), SP (top speed, miles per hour), and WT (vehicle weight in 100 lbs.).
a. Consider the following model:
MPGi = β1 + β2 SPi + β3 HPi + β4 WTi + μi
Estimate the parameters of this model and interpret the results. Do they make economic sense?
Solution:
S.No. MPG SP HP VOL WT S.No. MPG SP HP VOL WT S.No. MPG SP HP VOL WT
As expected, MPG is positively related to HP and negatively related to speed and weight. All coefficients are significant statistically. b. Would you expect the error variance in the preceding to be heteroscedastic? Why? Solution: Since this is a cross-sectional data involving a diversity of cars, a prior one would expect heteroscedasticity.
c. Use the White test to find out if the error variance is heteroscedastic. Solution:
Regressing the squared residuals obtained from the model shown in (a) on the three regressors, their squared terms, and their cross-product terms, we obtain R2 = 0.502 nR2 = 81 × 0.502 = 40.662 The 5 percent critical chi-square value for 9 df is 16.919. Since calculated value of chi-square (nR2) is greater than critical value (16.919). So we reject Ho and conclude that there is heteroscedasticity. d. Obtain White’s heteroscedasticity-consistent standard errors and t values and compare your results with those obtained from OLS. Solution:
When you compare this result with the OLS results, you will find that the values of the estimated coefficients are the same, but their variances and standard errors are different. As you can see, the standard errors of all the estimated slope coefficients are higher under the White test procedure, hence |t| are lower, suggesting that OLS had underestimated the standard errors. This could all be due to heteroscedasticity. e. If heteroscedasticity is established, how would you transform the data so that in the transformed data the error variance is homoscedastic? Show the necessary calculations. Solution: There is no simple formula to determine the exact nature of heteroscedasticity in the present case. Perhaps one could make some simple assumptions and try various transformations. For example, if it is believed that the “culprit” variable is HP, and if we believe that the error variance is proportional to the square of HP, we could divide through by HP and see what happens. Of course, any other regressor is a likely candidate for transformation.
