Part 3

Question 1

Assumptions Classical linear regression model (3)

Answer 1

Residual is on average 0, linearly independent and normal

Question 2

Consequences of ignoring autocorrelation

Answer 2

Regular standard error estimates inappropriate, typically too
small.
R^2
is likely to be inflated relative to its “correct” value for
positively correlated residuals.

Question 3

Causes of autocorrelation. (2)

Answer 3

Omission of relevant variables, which are themselves
autocorrelated.
Model is misspecified by using an inappropriate functional
form.

Question 4

Issue with Multicollinearity and give solution

Answer 4

Explanatory variables are very highly correlated with
each other. Drop one of the collinear variables

Question 5

measurement error

Answer 5

measurement error in one or more of the explanatory variables
will violate the assumption that the explanatory variables are
non-stochastic

Question 6

1 ways measurement errors occur

Answer 6

Macroeconomic variables are almost always estimated
quantities (GDP, inflation, and so on), as is most information
contained in company accounts
– Sometimes we cannot observe or obtain data on a variable we
require and so we need to use a proxy variable - e.g. expected inflation

Question 8

Parameter stability meaning

Answer 8

Parameter stability is the idea that coefficients in models (and
therefore the relationships between variables) must remain stable
over time

Question 9

Why is it desirable to remove insignificant variables from a
regression?

Answer 9

Insignificant variables refer to variables in a statistical model that do not show a statistically significant relationship with the variable you are trying to predict or understand. This means the evidence is not strong enough to conclude that their effect is different from zero.

Leaving them in the regression uses up degrees of freedom

Question 10

Solutions for the presence of autocorrelation in the residuals
includes

Answer 10

1. Use HAC robust standard errors, such as Newey-West.
2. Use a dynamic model to account for the autoregressive
   tendencies in the data.

Question 11

Normal distribution properties (3)

Answer 11

A normal distribution has zero skew and
▶ a kurtosis of 3;
▶ Skewness and kurtosis are the (standardised) third and fourth
moments of a distribution.

Question 12

what do outliers have an
effect on?

Answer 12

: Outliers tend have a large effect on the OLS
estimate

Question 13

solutions for outliers (3)

Answer 13

Remove outliers (need good reason!)
▶ Use of dummies
▶ Robust regression