Panel data estimation with a large number of dummy variables

[Dino_Jensen]

Hi!

I am relatively new to Eviews, but I am working on a panel data model with a large number of dummy variables (125 in total) and I am wondering how I can include them in the Eviews estimation without having to write them all in the equation itself.

About the model:

I am doing some work on the topic of "Diversification discount" within the corporate finance area. As a little initial test I am trying to estimate a model with Price/Earnings as the dependent variable, a dummy variable for diversification and a number of firm characteristic variables (Profitability, Risk, Size, Growth opportunities and Leverage) as well as a large number of dummy variables to control for industry effects (125 in total). My panel data set consists of 310 listed UK firms for which I have observations from 2008 to 2012. Basically the model looks like this:

P/Eit = a + bDIVit + bSIZEit + bRISKit + bGROWTHit + bPROFit + bLEVit + bD_13it + …. bD951it + eit

The variables D_13 to D_951 (and the 123 other dummy variables) are meant to control for industry effects and represent 3-digit SIC industry classifications.

Now my question pertains to the easiest way of doing the estimation in Eviews without having to write the above equation in full with 125 dummy variables? Is there some kind of IS-formula or anything that can be applied (as I said I am relatively new to Eviews)?

If someone could provide sort of a step-by-step guide I would be very grateful! I have included my data set in this post and hope that it will make it easier to understand the model in question.

[EViews Glenn]

[Sorry to take so long to respond, but we've been out-of-town at a conference]

The following code runs the regression in question.

Code: Select all

wfopen sample_data.xlsx pagestruct firm @date(year) group d_grp d_* equation eq1.ls pe size risk growth prof lev d_grp

Here I read in the date and structure it as a firm-level panel (not necessary for what follows). Then I create a group object with all of the dummy variables using a wildcard expression. This group is then added into the equation in the next line.

Note that I had to leave the intercept out of the equation specification as the set of dummies span the entire sample. Alternately, you could include an intercept and use a drop to remove a particular dummy from the group.

Code: Select all

wfopen sample_data.xlsx pagestruct firm @date(year) group d_grp d_* d_grp.drop d_13 equation eq1.ls pe c size risk growth prof lev d_grp

Note the presence of C in this latter equation.