Ch 9 - Modelling asset return volatility: extensions

Question 1

Why does the vanilla GARCH model not account for the leverage effect?

Answer 1

It does not allow for negative returns to have a different impact on volatility tomorrow than positive returns. The GARCH model assumes that tomorrow’s volatility is quadratic in today’s residual, so the sign of today’s residual does not matter

Question 2

What are 2 models that account for the leverage effect?

Answer 2

GJR-GARCH/ Threshold GARCH, EGARCH

Question 3

One way of illustrating the difference between volatility models is via their _________

Answer 3

News impact curves

Question 4

Describe the EGARCH model

Answer 4

The exponential GARCH model captures the leverage effect and forgoes the need to impose conditions on the parameters of the GARCH model to ensure a positive volatility estimate

Question 5

Suggest a model that accounts for the positive relationship between return and risk

Answer 5

ARCH-M/ ARCH in mean model

Question 6

What is the most common use of a volatility model?

Answer 6

Out of sample forecasting of volatility

Question 7

In general, more ____________ models fare better with out of sample forecasts, while more ____ models lead to more accurate in sample predictions

Answer 7

Parsimonious, specific

Question 8

Name 4 methods of choosing a volatility model

Answer 8

Nested models
Information criteria (AIC, BIC, HQIC)
Statistical goodness of fit measure (MSE, standard error, QLike, R^2)
Economic goodness of fit measures