Page 1 of 1

Programming an estimation without coefficient?

Posted: Sat Aug 07, 2010 4:15 am
by metcats4life
Good morning,

First, I´m new in this forum and as english is not my mother tongue I want to
excuse in advance for potential spelling or grammatical errors.

For my diploma thesis I have to do an empirical study about Carry Trades, including a regression of a potential profability of CT.
I want to recreate a method I have found in the litrature but failed so far in implementing in Eviews7.
Maybe somebody in this forum can give me advise.

My problem is that I want to do the same regression as you can see in the attached image.
How do I have to enter this regression equation in the "estimate equation" input field?
If I do it with the ARCH model I always get an coefficient but I just want to get the intercept!

Image

Than you very much for your help!

Re: Programming an estimation without coefficient?

Posted: Mon Aug 09, 2010 2:01 am
by trubador
This is not an ARCH estimation. You can specify the model as a regular OLS. In the "Equation Estimation" window, select the "LS-Least Squares Estimation (NLS and ARMA)" in estimation settings, which is the default. Type your equation in the "Equation Specification". Depending on the name of your variables it should be something like:

Code: Select all

RH-R c
Statistics of alpha parameter (the constant in your equation) will be displayed in the output. For JB statistics, you have to select "View/Residual Diagnostics/Histogram-Normality Test" and "View/Residual Diagnostics/Heteroscedasticity Tests/ARCH" to obtain the ARCH statistics.

Re: Programming an estimation without coefficient?

Posted: Tue Aug 10, 2010 3:16 am
by metcats4life
That works!! :-)
Thank you so much for your help!