21. In the following regression, write the auxiliary regression used in the White
test for heteroskedasticity:
$$y_t = \beta_0 + \beta_1 x_{1t} + \beta_2 x_{2t} + u_t$$
22. What is the null and alternative hypothesis of the White test of heteroskedas-
ticity?
23. If the p-value of the F-test in the auxiliary regression is 0.67, what would
be your conclusion? Is there evidence or not of heteroskedasticity and under
which significance level?
24. If the p-value of the F-test in the auxiliary regression is 0.02, what would
be your conclusion? Is there evidence or not of heteroskedasticity and under
which significance level?
25. What is the main disadvantage of the White test for heteroskedasticity?
26. If we know that the variance of the error term can be expressed as:
$$var(u_t) = \sigma^2 x_{1t},$$
how can we modify the original regression $$y_t = \beta_0 + \beta_1 x_{1t} + \beta_2 x_{2t} + u_t$$, so
that the error term becomes homoscedastic? Write down the new modified
regression.
27. What is the name of this approach to correct for heteroskedasticity?
3