Panel data OLS and fixed effects
Posted: Mon Mar 19, 2012 10:42 am
Hi,
My name is Mathias and I am currently working on a panel dataset concerning the Belgian alcoholic beverages industry (2001 - 2010).
I've created 4 seperate dummy's:
- Beer_dummy = 1 when the enterprises are active in the beer and malt industry and dummy= 0 when they aren't.
- Others_dummy = 1 when he enterprises are active in the other alcoholic beverages industries and dummy=0 for all the enterprises who aren't.
- Foreign_dummy = 1 when the company is owned by a foreign MNO (>25% participation) and dummy=0 when it isn't owned by a foreign multinational.
- Participation_dummy = 1 for Belgian enterprises who have (>25% participation) in foreign firms, namely Belgian MNO and dummy=0 when they are just Belgian firms and only active in the home market.
POOLED OLS:
Now my problem is, that when I execute an OLS - estimation (ordinary least squares) and I include all these dummy's together with my other variabels 'value added' (dependent variable), employment and 'labourproductivity'. I can't obtain the constant term 'C' in my equation; I'll get the error mesage 'NEAR SINGULAR MATRIX'.
Only when I execute the equation 'vallue added' (dependend variable), C (contant term), employment and labourproductiviy togheter with 2 dummy - variabels, beerdummy and foreign dummy. Then I'll get an output.
- Does anyone now if the values of the other dummy's 'others' and 'participation' are automatically reflected in my constant term (C)?
FIXED EFFECTS & RANDOM EFFECTS
When I work further with my last OLS equation (including the dummy's 'beer and foreign' and without dummy's 'others' and 'participation'.
And I put my panel options --> effects specification:
- cross - section: fixed and period = fixed my equation doesn't work.
It only works when I insert 'period = fixed'. Is this a problem?.
The hausman test from my random effects: indicated that I would have to work further with fixed effects.
- But is it OK to work further with this equation and without the cross - section on fixed in the panel options.
My last question is how I have to interpret the number 5,2537E-109. I now it is very small but I don't know how to interpret it.
I would be very grateful if somebody could help me.
Thanks in advance
Mathias
My name is Mathias and I am currently working on a panel dataset concerning the Belgian alcoholic beverages industry (2001 - 2010).
I've created 4 seperate dummy's:
- Beer_dummy = 1 when the enterprises are active in the beer and malt industry and dummy= 0 when they aren't.
- Others_dummy = 1 when he enterprises are active in the other alcoholic beverages industries and dummy=0 for all the enterprises who aren't.
- Foreign_dummy = 1 when the company is owned by a foreign MNO (>25% participation) and dummy=0 when it isn't owned by a foreign multinational.
- Participation_dummy = 1 for Belgian enterprises who have (>25% participation) in foreign firms, namely Belgian MNO and dummy=0 when they are just Belgian firms and only active in the home market.
POOLED OLS:
Now my problem is, that when I execute an OLS - estimation (ordinary least squares) and I include all these dummy's together with my other variabels 'value added' (dependent variable), employment and 'labourproductivity'. I can't obtain the constant term 'C' in my equation; I'll get the error mesage 'NEAR SINGULAR MATRIX'.
Only when I execute the equation 'vallue added' (dependend variable), C (contant term), employment and labourproductiviy togheter with 2 dummy - variabels, beerdummy and foreign dummy. Then I'll get an output.
- Does anyone now if the values of the other dummy's 'others' and 'participation' are automatically reflected in my constant term (C)?
FIXED EFFECTS & RANDOM EFFECTS
When I work further with my last OLS equation (including the dummy's 'beer and foreign' and without dummy's 'others' and 'participation'.
And I put my panel options --> effects specification:
- cross - section: fixed and period = fixed my equation doesn't work.
It only works when I insert 'period = fixed'. Is this a problem?.
The hausman test from my random effects: indicated that I would have to work further with fixed effects.
- But is it OK to work further with this equation and without the cross - section on fixed in the panel options.
My last question is how I have to interpret the number 5,2537E-109. I now it is very small but I don't know how to interpret it.
I would be very grateful if somebody could help me.
Thanks in advance
Mathias