OLS Slope When the True Relationship Is Non-Linear

Reported in interviews at Citadel

Problem statement

You run an OLS regression of $Y$ on $X$. What is the expected value of the slope coefficient $\hat{\beta}$ if the true relationship is:

(a) $Y = X^2 + \epsilon$

(b) $Y = X^3 + \epsilon$

Assume $X \sim N(0,1)$ and $\epsilon \sim N(0,1)$, independent of $X$.

Show hint

What population quantity does an OLS slope estimate when the true relationship is not a straight line?

Relevant roles: Data Scientist, Applied Scientist and ML Engineer.

Related free problems