GARCH(1,1) to estimate volatility

[Clickerz86]

Great video extremely helpful. Can I ask you a question if you don't mind? If one were trying simply to get the conditional variance of a set of data (monthly stock index returns in my case), to test whether certain months are positively correlated with said conditional variance, what is the best way to calculate this conditional variance? Thanks in advance.

[roshjesh]

this was really helpful ....i was wandering if i could get some details to add your name in my reference section..

[JSA19882007]

why do we add weights?

[JSA19882007]

hey, could you please send me the excel worksheet for using the garch (1,1) model please? the excel worksheet that i got is abit complicated, yours is simple.

[conflozed]

hi!- nice video! can i ask a little question?

How do you claculate variance n-1 (cell F13) is calculated as the variance(F14:F15) ?

Thanks

Flo

[Alveuz]

Does VIX from CBOE can be compared with Garch(1,1)?

If so, which has a better performance in calculate market volatility?

Thanks in advance

[mammutpenthouse]

@Alveuz Yes, they are both measures of market volatility, but remember to treat VIX carefully as it sometimes fails to do its cause, i.e. it doesn't always look forward as it implies.

[Alveuz]

Does VIX from CBOE can be compared with Garch(1,1)?

If so, which has a better performance in calculate market volatility?

Thanks in advance

[oringent]

Why are you using log returns? this is not a continuous but a discrete model.

[sangitawarade]

Hello Sir How to caluculate the value of alpha, beta and gamma???

[oringent]

you can OLS to estimate these parameters

[abhishekpreetam]

how can we predict the values of alpha, beta and gamma for our calculations. please tell the maths behind this so that it is easier for me to go ahead. :) waiting for ur reply :)

[rabbeborchert]

What does Column E mean? what do you measure with these weights?

[Napsugi04]

You don't really measure anything. You weight the returns with column E. You want to give higher weight to more recent observations. Instead of using 1/n equal weights to all of them.

[Napsugi04]

Weights column (E) seems to be wrong. You need to shift the cells down by one, i.e the 10% should be in line 14, otherwise the alpha = 6%, beta = 94% will not give you the same result as the EWMA!!

[nikiforosc]

For GARCH(1,1) estimation, Hull in his textbook, defines "u" as arithmetic return instead of logarithmic return. I noticed that in your calculations you used log return, which one we should use for our calculations?

[DannyChan333]

can I ask you one question that values of gamma alpha beta are all arbitrarily set by you with the condition that sum of them is 1. but could we just estimate the gamma alpha and beta by the software itself say, STATA or Eview?

[bionicturtledotcom]

Hi Danny, yes absolutely. Common is to use maximum likelihood method (MLE) to find parameters (weights) based on a sample. Thanks, David

[brohrox]

Thanks.. very simplified..Great job!

[mjmcgiven]

how do you work out the weightings of y, alpha and beta??

[armjay]

historical data from Yahoo finance

[kermanparts]

well explained.