Explanation

In multiple least squares regression, one of the key classical assumptions is that the independent variables are not perfectly multicollinear. This means:

• No independent variable can be expressed as a perfect linear combination of other independent variables
• The design matrix (matrix of independent variables) must have full column rank
• This ensures that the regression coefficients can be uniquely estimated

Why the other options are incorrect:

• A: Stationarity is not a required assumption for multiple least squares regression. While stationarity is important in time series analysis, cross-sectional regression models don't require stationary dependent variables.

• C: Heteroskedastic error terms actually violate the classical assumptions of OLS. The correct assumption is that error terms are homoskedastic (constant variance).

• D: Homoskedasticity refers to the error terms, not the independent variables. The assumption is that the error terms are homoskedastic, meaning they have constant variance across all observations.

Key OLS Assumptions:

1. Linearity in parameters
2. Random sampling
3. No perfect multicollinearity
4. Zero conditional mean (E(ε|X) = 0)
5. Homoskedasticity (constant variance of errors)
6. Normality of errors (for inference)

Option B correctly identifies one of these fundamental assumptions.