UCSD - GARCH Package of Kevin Sheppard. Multivariate GARCH models

Hello! I'm trying to use MV-GARCH model from UCSD package but I always have errors that are related to the data type.

For example : ??? Undefined function or method 'dcc_mvgarch' for input arguments of type 'double'.

I don't know a lot in MATLAB so can somebody explain to me how to change the data type and why it doesn't work?

As a data I have a 5000x10 matrix of residuals of rt=mu+et model. Is it ok?

Can somebody just show me a little example of how it works?

Thank you for all your help!

Answer 1

[Oleg Komarov]

You have to add the the folders (and subfolders) where the package is to the MATLAB path:

File > Set path > Add with subfolders

Answer 2

[Maria]

Hi Oleg! Thank you , it works much better but still there is a problem. I mean in the USCD toolbox distribution functions have their previous names like chis_cdf.m while I have rather new MATLAB version where there is only chi2cdf.m .

I tried to find old functions looking in the Internet but it didn't lead to the good results. So what is in your opinion the fastest way to solve the problem? For example I can't find mprint function in thee new MATLAB. What is the analog?

Thanks a lot!

[Oleg Komarov]

There's no guarantee MATLAB implements user defined functions. Also, no sure method to find potential analogs.

Try searching the FEX for other user defined functions that may suit your needs.

Answer 3

[Maria]

Ok , thank you. I'll try this way.

By the way Oleg , do you whether there are some other Multivariate GARCH toolboxes for MATLAB? I've found only the Sheppard's one.

Thank you!

[Oleg Komarov]

No, I dont know any other.

Answer 4

[Garth Mortensen]

My thesis applied MVGARCH to two return series. I put all my work and code online so others can benefit.

https://sites.google.com/site/garthmortensenthesis/

I hope it helps.

Answer 5

[Abdelrazzaq]

Dear all; In order to estimate the ADCC, can anybody please tell me the order of the functions in need to call to estimaate the model? please help

Is it enough to call the function DCC to get the dynamic correlation, if so then which output will give me the dynamic corr series, so i can use further in my regression