EViews.com

Hello,

I am writing my dissertation on mutual funds and whether returns are affected by their position with a financial centre (modeled by a dummy variable). I wish first to use a fama-french model:

r_i - r_f = alpha +b1exrm + b2hml + b3smb + e_i

to adjust returns for risk. Then I will regress these returns on the dummy variable. However I am unsure how to convert the data from excel to eviews. I have provided a sample of the data (for one of my time period and those with +1 dummy variable) as an example of how my work is setup. Please can someone advise me on how to proceed?

Many thanks, Dan

Drag-and-drop your Excel file onto EViews?

Yes I have tried doing it like this but I am getting a collinearity problem when I run the regression. I think there is a problem with the way I am importing it. I have tried importing the whole sample as panel data but it didn't seem correct. Now I am splitting the whole sample into smaller samples with the same date, so it is more like a cross-section.

You might want to post your EViews workfile, including the equation that is giving you a problem.

Good idea. Here it is. The equation I used was excess_return_fc c mkt_rf_fc smb_fc hml_fc mom_fc

Have you looked at the data in EViews and seen whether it matches the data in Excel? If it matches, the way you imported the data was not wrong :D

Have you looked at your data in EViews?

All your right hand side variables are constants.

They are constant because they are the same variable for each observation. They should only vary over time, but I have taken a cross-section here so the time is not changing. Would you recommend using a different method?

Yes, use a time series or panel. The cross-section regression is impossible to estimate. This has nothing to do with EViews.

Okay, you're right actually I think it works fine in the panel data setup. I now have the risk adjusted returns. Would you be able to advise me on how to regress these on a dummy variable. Do I need to create a eviews file like the 'dv' one I have here where it looks like 1,0,0,1,0,1 sort of format?

Load your data into EViews and then create a dummy variable inside the EViews workfile. Then use that dummy in the regression as you would any other variable.