Hello Experts! I'm conducting a study on the effect of trade openness on industry competition. My panel dataset is composed by 3 dimensions: country (5), industry (19) and time (9 years). The purpose of my study is to capture a dynamic equilibrium relationship. For that I need to distinguish between the short-run impact of a trade openness on industry price markups and the long-run equilibrium relationship between price markups and the competitive environment. It is important from an econometric point of view to control for the high degree of persistence of price markups: an Error Correction Model specification I allow industries to trend to different equilibria.
Now my question concerns the stata code for an error correction model with a 3 dimension panel data. I'm also indecisive between combining country and industry information to one panel unit instead of having 2 units - which doesn't let me perform xtset due to repeated time values within panel.
Thank you for you attention and time to consider my issue!
Would it be wrong to just use
xtset country time
and then use the industry dimension as a dummy variables which varies over country and time?
Jakob, since I have the data in long form (country industry and year) and then columns with the variables, doing xtset country year gives me the error message "repeated time values within panel r(451);" . But concerning the error correction model for panel data, do you know which one is the most appropriate?
Concerning the repeated time values within a panel I would recommend to check http://www.stata.com/support/faqs/data-management/duplicate-observations/
Depending on from where you got the data, maybe there is just a mistake in the data you could find using the described method.
To primary question: Can't help you much, but I would recommend the following papers:
"Error-correction–based cointegration tests for panel data", The Stata Journal (2008), 8, Number 2, pp. 232–241
Blackburne, E. F. and Frank, M. W. (2007), Estimation of Nonstationary Heterogeneous Panels, The Stata Journal, Vol. 7 (2), pp. 197-208
Maybe those papers can help. A stata command which is often named with the ECM is the xtwest. But I have little experience with it.
Here is the answer, I would reccomend you first do the principle component analysis then dimension reduction (factor analysis) thats why zou would have less data in your hand which are much more significant for you. Later on do the Cluster analysis. then you will have the result.
Sincerely
Hakan
Dear Torrinha!
I can help you in E-Views command bcz i am an E-Views user, if u are easy with E-Views.
Thank you
Thank you all for your answers!
For my research I ended up resorting to the Dynamic Panel Data models and performed a comparison between the Pooled Mean Group (PMG) estimator, the Mean Group (MG) estimator and the Dynamic Fixed effects (DFE). The most intermediate alternative between these three, and thus, the one I ended up interpreting the results from, is the PMG estimator.
In order to estimate the vector error correction model with panel data I used the Stata-based software developed by Blackburne and Frank (2007) (command xtpmg). Thank you for the tip Jakob Schwerter !
Hope this helps others with the same doubt! I can also answer further questions about this model ( I really dived into it during this research).
This link should be of help to you
https://www.princeton.edu/~otorres/Panel101.pdf
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size