I am trying to run a regression of cobb douglas function:
The problem that my dataset capture the firm at a point of time,
So I have a dataset over the period 1988-2012.
Each firm appears one time!
(I cannot define if it is a panel/time series/cross section..)
I want to find the effect of labor, capital on value added.
I have information on intermediate input.
I use two methods Olley& pakes, levinsohn-patrin.
But Stata is always telling me that there is no observations!
my command:
levpet lvalue, free(labour) proxy(intermediate_input) capital(capital) valueadded reps(250)
Why the command is not working and telling that there is no observations?
(Is this due the fact that each firm appear only one time in the data?)
(If yes, what is the possible corrections for simultanety and selection bias in this data?)
Thanks in advance for your help,
Mina
My experience of production function estimation predates OlleyPakes and Levinson-Patrin. However it can be stated that you can not use panel estimation methods (either fixed or random firm effects) if you have only one observation for each firm. This is why your stata panel estimation will not work
It may be possible to stack your observations adding either dummy or trend variables for the date (What you do here depends on economic reasoning). As you will probably have endogeneity problems you may need to use instrumental variables or estimate a simultaneous production system of its dual. This of course assumes that the production function is the same across firms and over time. You have an interesting project.
Dear Mina,
If you have enough data on each firm, then you can do time series analysis for each individually.
You can still do cross-section analysis with multiply firms. I attach a description of a classical study by P. Dhrymes that was done on cross section using system of equation analysis. Good luck in your project.
Dear Mina,
If your command is correct, then the cause of the non-functioning could be that your dataset is not well-organized. Why do you not try simple OLS? As I understand you have the following data vectors: name of firm, year (T), inputs (I), labour (L), capital (K), value added (VA). I see no statistical-technical problem to estimate, for example, a Cobb-Douglas function like VA=a*T+b*lnI+c*lnL+d*lnC+const (where I think that the inclusion of I is problematic, because VA is simply added to I to get gross production..
The main problem is not the estimation, but the economic assumptions you need and data adjustments you may need for a sensible estimation. As John writes there is the assumption that all firms have the same production function (not necessarily over time, but the same technical progress. This fits to the modern models with the representative firm, but I fear that it is not realistic.There are also numerous measurement problems.
Correction and supplement
In the CD-PF it should, of course, be lnVA=....
If you know which industries the firms belong to and if there are sufficient data, you can give up the assumption of only one production function and estimate PF for several industries.
Hello Mina Sami , I believe you have a "pool of cross-section data". Then you would only be able to use cross-section or pseudo-panel techniques and some time controls and instrumental variable. Sugestion: pseudo-panel techniques
- Badges
- Science topic
- Similar topics
- Biological Science
- Anatomy
- Tissue