Part 3 - part 2

Question 1

With BP, White Test and Arch-LM, if we reject the null what do we conclude?

Answer 1

hetereoskedasticity

Question 2

ARCH-LM assumption, problem and solution

Answer 2

▶ Assumption: residuals are linearly independent.
▶ Problem: If violated, standard errors are off.
▶ Solution: Robust standard errors such as Newey-West/White

Question 3

Conditions which Must be Fulfilled for DW(type of autocorrelation test) to be a Valid Test (2)

Answer 3

Conditions which Must be Fulfilled for DW to be a Valid Test
1. Constant term in regression

2. No lags of dependent variable to assess the independence of residuals

Question 4

Breusch-Godfrey Test, rejection implies …

Answer 4

rejection implies presence of
autocorrelation.

Question 5

(E(ϵ) ̸= 0 problem, what is solution

Answer 5

Include intercept

Question 6

Heteroskedasticity - what is solution

Answer 6

Robust Standard Errors

Question 7

Residual is correlated - what is solution

Answer 7

Robust Standard Errors / Model Dynamics

Question 8

Var. not exogenous - what is solution

Answer 8

Instrumental Variables

Question 9

5. Resid is not normal - what is solution

Answer 9

Deal with outliers

Question 10

Chow Test Formula

Answer 10

test statistic =
RSS − (RSS1 + RSS2) /
RSS1 + RSS2
×
T − 2k
k

Question 11

Chow test needs enough data to do the regression in both
sub-samples. What is the alternative (2)

Answer 11

the predictive failure test.
▶ Estimate the regression over a “long” sub-period (i.e. most of
the data) and then we predict values for the other period and
compare the two.

Question 12

Forward predictive failure test

Answer 12

Forward predictive failure tests, last few observations used for
forecast testing

Question 13

Backward predictive failure test

Answer 13

▶ Backward predictive failure tests, “back-cast” the first few
observations

Question 14

If Chow test null is rejected conclude that …

Answer 14

parameters are not stable