Cross Sectional Analysis

[indigow]

Hello All,
I am new to eviews and need to do some analysis for my homework. I hope I am creating the topic to the right place and glad if someone can answer it.
I am working on an APT model, which has 4 variables. I was able to identify the significant variables with time series regression but as a second step, I need to run a cross sectional regression:
'However, CRR used a version of the Fama-MacBeth (1973, 1974) technique which consists of a two-stage regression. The first set of regressions estimates the portfolios' exposures to pricing factors (betas). The second set of regressions estimates the market prices for the beta values obtained from the first set of regressions.The result of this two-stage regression methodology is to generate time series of estimated premia for each risk factor.'
I should be using the mean returns and regress it with the coefficients found in time series regression; but can someone explain how I might do that?
It might be easy question (I am sure it is!) but I would be glad if you can provide how I might achieve it.

Thanks in advance!

[EViews Gareth]

Why not use the Fama-Macbeth add-in?
http://www.eviews.com/Addins/addins.shtml

[indigow]

Thanks for the reply Gareth. I am not sure if I would be able to install an add-in though, since the one I use belongs to school.
In theory, I should be able to do the cross-sectional analysis, with the coefficients I got from the time-series regression and regress it with average returns (which is my dependent variable for the time series analysis), but how?
Should be simple I suppose, but I could not find anything about it online.

[EViews Gareth]

Copy the coefficients into a new undated page, convert them into series objects, the run the regression

[indigow]

Yes! but I need some more details. What I did was:
I calculated the average return for 10 companies. (series 1)
I took the coefficients I found from the time series regression for each variable into arrays.
So in total I got 5 series.
Can you please confirm this is the right way to do it?
After that, I did Quick/estimate and wrote "series1 c series2,3,4,5", but again, since the time period is still down there when I estimate the equation, I am not sure if I am doing it right or wrong. Should I use another tool for the cross sectional analysis?
I hope what I wrote above is clear and glad if someone can clarify what I should be doing. any 'how to" file would be appreciated as well!

[EViews Gareth]

How did you get 5 series?

Explain in detail exactly what you have done.