Panel Data Analysis: A Complete Guide

This Lecture deals with panel data analysis which covers descriptive statistics, correlation analysis, multicollinearity diagnostic tests through (VIF), regression analysis through static panel (fixed effect, random effect, Hausman test and their decision criteria) & dynamic panel data analysis (GMM) along with commands of these all tests. Furthermore, it explains how to report results in Excel and Word documents" Following are the commands:
Stata Commands for Panel Data Analysis
1) Import data:
Method 1:
File ____ Import ______ Select first option ______Browse _______ Import first row_____Enter
Method 2:
Select excel file _______ Enter “edit” in Command box ________paste file and then minimize the file
2) To generate panel:
xtset id year ___enter
3) Descriptive statistics
summarize ______enter _________select table, right click and select copy table, then open excel sheet, paste special and click on text and then Ok. Delete extra rows
4) For correlation
pwcorr ________enter
5) For vif (run simple regression)
reg DV IDVs CVs and then enter
vif and enter
6) For Static panel regression
Step 1: xtreg DV IDVs CVs, fe _________enter
Step 2: xtreg DV IDVs CVs, re _________enter
Step 3: estimate store fixed ______ enter
Step 4: estimate store Random _____ enter
Step 5: hausman fixed Random _______ enter
Note: If p-value of Chi square is less than 0.05, select fixed effect results, otherwise random effect results.
7) For Dynamic Panel (GMM)
xtdpdsys DV IVs CVs, lags(2) maxlags(2) twostep artests(2) _______ enter
For Sargan
estat sargan _____ enter (p-value should be insignificant to validate instruments)
For AR1 and AR2
estat abond _____ enter (AR1 significant and AR2 insignificant)

[@kentiy1]

This video is very usefull and informative, thank you, Dr. Hunjra

[@samiurrahman979]

Dr sb. thank you for providing this information. This video is very informative. However, in the next video, kindly discuss the nature of static and dynamic data, auto-correlation problems or all basic assumptions in the data and treating such issues in the panel data. we are waiting for your next lecture.

[@farhaanrashid8852]

What if some independent variables are time series( Macroeconomic variables)and some are panel ( Micro economic variables). ?

[@fadeldiande7658]

Ma sha Allah, good job. Sir, I would like to know if we should do stationary test and cointegration test before starting the regression ?

[@tanveerbagh8524]

Very informative lecture.

[@aihunjra]

Thanks for liking

[@user-kr4pm3wq9x]

There is a very major mistake in hausman test You run fixed, then random then you stored two times, actually you store random results two times(you can see the numbers are same) you need to run fixed, store, run random and store, then hausman

[@muhammadkhalidsohail2994]

Masha Allah Dr. sb. keep it up

[@aihunjra]

Thanks a lot

[@munshir.c6161]

thank you so much

[@munshir.c6161]

i have one doubt sir: when we go for GMM test? doest it go when Fixed effect model is not appropriate? if it is so , how we can know fixed effect model is not appropriate? kindly, Thank you

[@aihunjra]

Dear, as I understand, one is the static panel and the other is the dynamic panel approach which is superior to static and the results are more robust and efficient, we normally check robustness of results by applying the both

[@aihunjra]

thanks dear

[@abdisheycho564]

good jod dr.

[@mariafarheen8387]

aoa sir, what if in GMM, p values are insignificant ?

[@samiraabdous2058]

Could you send or out the data here?

[@alishbarahmanullah9144]

Sir, can you make one video for series data too for stata software. Thank you sir.

[@aihunjra]

Dear it's in plan, soon I will

[@alishbarahmanullah9144]

@@aihunjra thank you sir. Will be waiting sir.

[@drkeyurnayak7812]

share datafile

[@dhananjayak1095]

Your chi square value in Hausman test is 0 .. It is not the p value.. There is some problem as you didnt get p value.. So, your interpretation is wrong Try hausman fixed random, sigmamore