. Imagine we fit a linear model where yi is real GDP per capita in the United St

Question

. Imagine we fit a linear model where yi is real GDP per capita in the United States in each year from 1900 through 2016. We believe that citizen education is the best predictor of GDP, so we fit a model yi = 0+1xi +²i where xi is the percent of Americans with a college degree in year i. Other than omitted variables, which if any Gauss-Markov assumptions are we most likely to have violated?

a. Normality of errors

b. Parametric linearity

c. Constant error variance (homoskedasticity)

d. No error autocorrelation

e. None of these / cannot say

Explanation / Answer

since xi depend on year , we can not have constant variance

hence

c. Constant error variance (homoskedasticity) is violated

Related Questions

Navigate

• Browse (All)
• Browse #
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.