Volatility Models Flashcards

1
Q

What does evaluation of volatility forecasts differ from regular forecast evaluation?

A

We do not observe realisation of variance/volatility, hence we need a proxy to evaluate the forecast

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What are typical variance proxies for forecast evaluation?

A

Squared returns

Squared errors

Realized variance

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is a Diebold Mariano for volatility forecasting?

A

A test of relative performance of 2 forecasts. Typical loss function is MSE (again need proxy) or QML statistic. Always estimate errors with Newey West. DM statistic is distributed N(0,1)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What are some problems with Realized Variance?

A

1) Data quality (typos, non-audited) 2) Market closure 3) Only observed in discrete grid 4) Price data not available at all times (can use interpolation) 5) Bid-Ask Bounce (causes upwards bias)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What are some solutions to the problems of Bid-Ask Bounce in Realized Variance estimation?

A

Less frequent sampling Remove bounce using MA(1) Use Mid-Quotes Use RV-AC1 estimator (unbiased yet not persistent)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What is the RV-AC1 estimator

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How to solve the problem of Market Closure?

A

Add squared close-to-open return. Potentially weight it in, using an appropriate weight

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is a Volatility Signature Plot?

A

Plots average RV at different sampling frequencies. Pick a sampling frequency where RV is stable. Too high and RV is upwards biased from bid-ask bounce, too low and it underestimates variance.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How can variance be modelled using RV?

A

Either directly using an ARMA or as latent using normal ARCH-family models

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What is a HAR model

A

A restricted AR model on Realized variance. Captures daily, weekly and monthly patterns. Regresses RV_t on RV_(t-1) and average RV for 5 first lags and RV for 22 first lags

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How to use RV as latent variable?

A

inse

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the VIX

A

It is a model-free measure of implied volatility. It uses both calls and puts, mainly out-of-the-money for higher liquidity

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What are some problems with model-free implied volatility?

A

Only finite number of calls Thin trading (liquidity) Needs to discretize the integral

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Write out an ARCH(1)

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

White out an GARCH(1,1)

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Write out an GJR-GARCH(1,1,1)

A
17
Q

Write out a TARCH(1,1,1)

A
18
Q

Write out an EGARCH(1,1,1)

A
19
Q

Write out an APARCH(1,1,1)

A
20
Q

Which types of options are used in VIX?

A

Only out of the money (for higher liquidity). Both call and puts

21
Q

How do you scale weekly volatility to annual?

A

sqrt(52) * weekly vol

22
Q

What is the general for of the Mincer Zarnowitz for forecast evaluation?

A