10 Heteroscedasticity
In cross-section data the spread of the dependent variable usually changes with the size of the unit. Large firms’ investment varies more than small firms’; rich households’ spending on food varies more than poor households’; the total income of a populous state varies more than that of a small one. The disturbances then violate assumption (A2), and \(\Var (u_i) = \sigma _i^2\) differs across observations. This section works out what that does to least squares, how to estimate standard errors that remain valid, how to detect heteroscedasticity, and how to exploit it for more efficient estimation.
10.1 Consequences for least squares
10.2 Heteroscedasticity-robust standard errors
10.3 Testing for heteroscedasticity
10.4 Estimation under heteroscedasticity
10.5 Heteroscedasticity in software
10.6 Practice Problems
10.2 Heteroscedasticity-robust standard errors
10.3 Testing for heteroscedasticity
10.4 Estimation under heteroscedasticity
10.5 Heteroscedasticity in software
10.6 Practice Problems
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.