Hi,
I am trying to examine the impact of monetary policy on net margins for 3-digit manufacturing industry using quarterly data. So basically, across rows I have quarterly data and across columns the net margins for 3-digit industry. One of the crucial reasons why SURE model is used is that since the impact of interest rates on 3-digit manufacturing industry net margins has been estimated, the error terms across the industries tend to be correlated due to the fact that the characteristics across the industries may be related. Also the RHS variables are all same across equations.
Now there are several questions which I am having while using SURE. Some of them are as follows:
- Since its a panel data, should I run sureg or xtsur command in STATA?
- How do I account for bias introduced by lag dependent variable (net margins)
- Should I also account for industry and year fixed effects since it is a panel data model?
Thanks
Hi Parul,
I think you can use xtsur command for random effects estimation. Another alternative is to use the command sureg including i.year and i.industry.
In order to account for lagged dependent variable it is better to use dynamic panel method. Please find the link below for details.
http://statisticalhorizons.com/lagged-dependent-variables
Joyita
Dear Parul,
when one deals with panel data, one should care much about the model and the estimation methods (estimators) he wants to use. So think about this!
Just simple answers to your questions:
1. it's obvious to use xt commands as along as you have a panel dataset (cross section time series)
2. there are many panel data methods including IV estimation when lagged dependent variables are regressors.. estimators are different depending on type of panel data you have..... when you have a long or short panel and mixed models... unbalanced or balanced.. anyways you need to specify what type of panel data do you have for you to choose the suitable model and estimator
3. it's always good to account for individual and time fixed effects when having a panel data
For more details on models and estimators you can read Cameron trivedi (2010) Microeconometrics Using Stata, Revised edition and Greene (2012) econometric analysis 7th ed.
Good luck
Yvonne
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series