turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Interpret the output of the OUTHT in proc varmax, ...

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

09-02-2016 08:45 PM - edited 09-02-2016 08:48 PM

I try to understand the output table generated by the OUTHT option in proc varmax, garch statement. Can anyone help to confirm whether my following interpretation is correct or not? thanks a lot!

In the output table ht which is generated by the OUTHT option in proc varmax, garch statement:

for a bivariate case, there are variable y1 and y2.

(1)

"Conditional Cov h1_1" represents the conditional variance of y1;

"Conditional Cov h2_2" represents the conditional variance of y2;

"Conditional Cov h1_2" represents the conditional covariance of y1 and y2;

therefore, conditional correlation 'rho' can be calculated as: conditional rho = h1_2 / (sqrt(h1_1)*sqrt(h2_2));

(2)

according to the link http://support.sas.com/documentation/cdl/en/etsug/63348/HTML/default/viewer.htm#etsug_varmax_sect041...

"The OUTHT= data set contains prediction of the fitted GARCH model produced by the GARCH statement."

if my original data has 21 observations (or just say 21 days), the values in OUTHT table are conditional covariance for those 21 days? or for predictions based on those 21 days' estimation? or include both?

Thanks!

************************

*use DCC (1, 1) - GARCH (1, 1);

proc varmax data=sample method=ml outest=a1_est outcov outstat=a2_stat;

by id year month;

model y1 y2;

garch q=1 p=1 form=dcc outht=ht;

output lead=21 back=0 out=forecasts;

run;

proc varmax data=sample method=ml outest=a1_est outcov outstat=a2_stat;

by id year month;

model y1 y2;

garch q=1 p=1 form=dcc outht=ht;

output lead=21 back=0 out=forecasts;

run;