I'm considering the regression model $y_i = \beta_0 + \beta_1x_{1i} + \beta_2x_{2i} + \varepsilon_i$ where the $\varepsilon_i$ are iid and $\mathcal N(0,\sigma^2)$
A study question asks to show the relationship between the variance of the least square estimates of $\beta_i$ and ${\rm VIF}_i$ for $i = 1,2$.