You are estimating a cross-section regression for a sample of 100 cities in the

Question

You are estimating a cross-section regression for a sample of 100 cities in the US in which you hope to explain expenditure on education as a function of the median income in the community, the number of school-age children, and the level of state and federal grants received for educational purposes.

Would you expect hetroscedasticity to be a problem in this case? If, so what would be a proper way to test for hetroscedacisity? Would you use the Goldfeld-Quandt test or the White test? Why?

What would be the statistical properties of the estimated coefficients if you don’t do proper correction?

How would you correct for hetroscadecity?

Explanation / Answer

As this data is cross-sectional, we would definitely expect heteroskedasticity.

White Test would be the most suitable in this case as here we don't know that the error variance is related to which variable.

In the presence of heteroskedastic errors,

We can correct heteroskedasticity in the following ways:

Related Questions

Navigate

• Browse (All)
• Browse Y
• Subjects

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