Multicollinearity, what it does and does not do
Your regression includes two factors that are almost perfectly correlated (say, two closely related value signals). Coefficients swing wildly and flip sign when you add a few data points.
Does near-collinearity bias your estimates? What exactly does it damage? How is it different from perfect collinearity, and how do you handle it?
Your answer
This one is open-ended. Work it through, then check your reasoning against the full solution.