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

Showing results for

Find a Community

- Home
- /
- SAS Programming
- /
- SAS Procedures
- /
- Proc Varmax: Cointegration and Error Correction Te...

- 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
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-26-2016 09:30 AM

Dear all:

I am interested in using the Proc Varmax function in SAS for the Cointegrating tests and vector error error correction models :

However, I would like to get some advice as to the steps that I need to follow for the model specification in a simple manner.

__For Cointegration tests:__

- i guess the first step is the Augment Dickey Fuller test. What are the steps that I need to use from beginning to the end ?
- What is the right procedure for determinging the lag order specification? Is there a tool in SAS that makes the software recommend « an automatic lag lenght selection ?
- Moreover, in terms of deterministic test assumption of test, how can I determine which model to use (in terms of selecting « no intercept » ; trend ; « trend and intercept »).What is the most optimal way to determine this ?

__Error Correction Model__

- I need to obtain the coefficient for the error correction term. This is a figure that I expect it to be between -1 and 0. Where should I look to get this
__single figure__in the results. Where exactly should I look at. Obviously, this is included in the alpha estimates of the model ? Which element of the matrix is important.

- A second question is that I saw that the model sometimes normalizes the beta estimates. Does this normalization have any impact on the beta estimate ?

I thank you in advance for your support