Is EGARCH model is best for time series data (stock market volatility). why we use EGARCH model rather not using any other model like TGARCH , MGARCH,, QGARCH, GJR-GARCH, FGARCH, COGARCH etc. what is the basic difference between these models.?
Usually when we are new to a field, we follow te methodology of someone who has done the choosing for us. so before you start, read the work of someone before you and if the reason he chose the model is adequate enough for you go for it. But in any case study first, don't just go around wondering what the difference is. If you work on a model and study it well, then you will be able to tell why this model is better, different, augmented, or whatever the difference from any other model you are comparing it with.This is the academic but also efficient way of doing econometrics.
Hi Anil,
First of all I think (or should I say "believe" as well ;-) ) that there is no single model that can be named "best" and hold the title for all different situations. Extensions of GARCH are suited for different cases and have both advantages and drawbacks. It is up to you, as a researcher to select a model and justify this selection in a proper way (perhaps by comparing its results with other possible choices and at least by demonstrating (if proving is not possible) that a particular model provides better/more accurate results).
In the context you mentioned, TGARCH for example could be used to model asymmetric response from investors to positive and negative news in the market.
Best regards,
Stan
There is no globally best model. The EGARCH cannot always be the best. Neither can any other model. Each model has its merits and limitations. Find attached herewith a paper for which GARCH and TGARCH did better than EGARCH for modelling Indian stock exchange volatility. Cheers.
EGARCH and JGR GARCH used to captured asymmetry in data. If the data is non stationary then we go for FGARCH or fractional model.... normally GJR with t distribution is quite good.
Model selection is a function of your objectives & your data. No one model is best in all cases & one has to select a model that best fits given the particular case. While doing literature review, pay adequate attention to the reason why a particular model was chosen in the first place & could there be something better. If subsequent research has used variants of the model, were the results / interpretation/ explanation better? This would help immensely when you collect your own data to justify the selection of a particular model.
One of the main advantages of the EGARCH is that is models logarithm of volatility. Therefore, during the estimation, there is no need for parameter restrictions. On the contrary, when you are estimating e.g. simplest GARCH(1,1), it is common that alpha and beta are restricted by the estimation procedure to be larger than zero. This is desirable, but in the EGARCH model, no such a restriction is needed.
There is no better model, there is the model that best fits the data. If you use the ARCH and GARCH model you can compare the result with IC (information criteria).
Sincerely
Fabrizio
It might be useful to read the following short papers on EGARCH:
“A one line derivation of EGARCH”, Econometrics, 2(2), 2014, 92-97.
“The correct regularity condition and interpretation of asymmetry in EGARCH”, Economics Letters, 161, 2017, 52-55.
“On the invertibility of EGARCH(p,q)”, Econometric Reviews, 37(8), 2018, 824-849.
- Badges
- Science topic
- Similar topics
- Philosophy Of Science
- Models