Multicollinearity versus heteroskedasticity: what each one actually breaks
Two of your regressors are highly correlated with each other (for example two nearly identical value signals).
Does multicollinearity bias your coefficients? What does it break, and would robust standard errors help? Contrast it with heteroskedasticity.
Show a hint
Heteroskedasticity leaves the true variance right and only breaks the estimate of it. Multicollinearity does the opposite: which one inflates the true variance of the coefficients?
Your answer
This one is open-ended. Work it through, then check your reasoning against the full solution.