Page 1 of 1

GARCH(1,1) Dummy Interpretation

Posted: Sat Jan 02, 2021 8:40 am
by MRocha98
Hello everyone,

I'm a beginner in time series. I'm trying to measure the impact of COVID-19 on the UK stock market volatility. I have the log daily returns (log_dr) and a dummy variable (c1) that assumes the value of 1 after the day 29 January 2020 which was the day that the UK registered the first cases. I'm using the model GARCH(1,1). What I did until now was:

1. Import my excel with the dates, the adj close, the daily log returns, and also the dummy c1
2. Select "Quick" and then "Estimate Equation..."
3. Then I selected the Method Arch
4. Mean Equation fill: log_dr ar(1)
5. Variance regressors: c1
6. Error distribution: Normal(Gaussian)

After this, I don't know how to interpret the coefficient of the dummy variable. I know that if it was in the mean equation it would mean that during the period of C1 we would get an increase/decrease relative to the non C1 period as great as the coefficient, however, I'm not sure if the same interpretation applies when one uses a dummy variable of the conditional variance equation.

Captura de ecrã 2020-12-29, às 16.00.39.png
Captura de ecrã 2020-12-29, às 16.00.39.png (264.73 KiB) Viewed 1014 times


I would also like to ask if someone could elaborate on the difference between conditional and unconditional variance and how relevant that difference may be to impact the possibility of assessing the impact on volatility I aim to study.

Thank you all for helping me!
Best regards!