Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: Linear DPD estimation (Dynamic Panel Models)
From
Mine Aysen Doyran <[email protected]>
To
[email protected]
Subject
st: Linear DPD estimation (Dynamic Panel Models)
Date
Thu, 07 Nov 2013 10:29:47 -0500
Dear list,
My apologies if this is rather a basic question. I am new to Linear DPD estimation. I hope somebody can help me.
First of all,
1) Do I have to do any unit root testing before I include any variables in dynamic panel models? (to see whether they are first difference stationary, for example)
2) In the diaolog box, there are three commands regarding the type variables included in estimation. Here are my questions:
2a) Dependent variable: I put my Y here (in level form, no lag)
2b) Independent variables: Does this section include all "endogenous" and "exogenous" variables of the model? ( I assume both).
2c) Gmm-Type Instruments (difference equation required*): What variables do I exactly include here? "Lagged dependent" variable? All endogenous variables of the model? or any endogenous variables "not" in model? any exogenous variables? or both? For the time being, I *only* put the lagged dependent variable.
2d) Standards instruments: What do I exactly include here? I am following Wooldridge's definition of IV Any variable that does not show in the equation but "correlated with the endogenous explanatory variables"? Is it desirable to follow this definition?
how can I determine the best "instrument"? Is it wise to run pairwise correlation and then chose variables partially correlated with "endogenous explanatory variables"?
iv: Instrumental Variable: “In an equation with endogenous explanatory variable, an IV is a variable THAT DOES NOT APPEAR in the equation, is uncorrelated with the error in the equation and is partially correlated with the endogenous explanatory variables” (Wooldridge, Introductory Econometrics, p.850).
2e) How do I conduct Hansen test of overall validity of instruments? I know the command for Sargan (estat sargan) but there is no command for hansen.
2f) How do I conduct serial correlation test of AR 1, AR2? It does not report anything after running estat abond
I really appreciate any help in specifying the variables under each category and information for robustness tests.
best regards,
Aysen
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/