Panel Regression

[Tarun Kumar]

Hi,
I am doing a research paper on panel data taken from BRICS nations as representative of recently industrialized and emerging market developing economies . i am using stata
After declaring my dataset as panel data, I have appointed general pooled ols model to do the regression which gave me R-squared as only 0.7421. intuiting it as not sufficient i went to fixed effect(fe) model, but it further reduced my (overall) R-squared to 0.1251 . then, i did random effect (re) model which gave (overall) R-squared as 0.7421. then, i did hausman test to choose between fe and re, which gave chi-squared as negative to be -205.32 and thus, hausman failed

My questions:
Q.1) how to choose between pooled ols and fixed effect? Please describe command too, if possible.
Q.2) as hausman test failed, how can i choose between fe and re? Please describe command too, if possible.
Q3) does any other model exist that can be used for panel regression other than pooled ols, fe and re in general? does any of them can be used for my study? If yes, how to choose between them?
Q.4) what kind of other test can i employ in my study to check characteristics of my data?
Q5) does things like heteroscedasticity, unit root test, autocorrelation etc creates a problem in panel data? if it does, how can i check them?
Q.6) any other suggestion for my study is helpful and so highly appreciated.
Q7) I am log of variable to signify growth. am i right? if not, when do we take ln of dependent and independent variables?
Q8) does 72 observations sufficient for my study on government spending and unemployment?

thank you for any and all kind of help that canbe provided for the study.

Ps.1) please excuse my poor knowledge about econometrics. please answer as soon as possible as i have a close deadline to submit my work
2) i am extremely sorry if i asked some elementary questions which u may find not worth answering. in such a case, plese tell me the question number and if possible, the URL link where i can read about it.
3) Please excuse length of the post. i know some may find it lengthy but to provide every related information, i choose this structure.

[Carlo Lazzaro]

Tarun:
welcome to this forum.
Unfortunately, this is by no means the way queries shoud be posted (see the FAQ on this and other posting-related topics).
Besides, your questions look boldly related to a home/classwork, something that is not welcomed (at least, not in the way you structured it, with no attempt to sketch something by yourself) on this forum (as per FAQ again).
It is really hard to believe that you teacher/professor/supervisor assigned you such an unfair challenge, given your declared poor knowledge of even the basic building blocks of panel data econometrics (this is something pretty easy to fix: just study any decent panel data econometric textbooks for Stata users: https://www.stata.com/bookstore/micr...metrics-stata/).
Eventually, urgency/deadline are something every Stata listers (and many other people all around the world) face daily and cannot be considered a voucher to obtain a fast and helpful reply, also considering that this is not a money vs consultancy hotline.

[Tarun Kumar]

Thank you for the reply Cario.
Instead of homework, this actually relates to my masters dissertation so instead of any teachers assignment, i took the task myself. and these questions are also actually written by myself. i framed it in this way to give all the relevent information so that i can get appropriate help and by doing so, didn`t realized the fact that i constructed it into a homework question. but since, as i have already posted and you given a look at it (for which i am thankful), i am going to leave it as it is. but i promise to take care in further posts.
As you have stated, I might have asked some basic elementary question but some of them are primarily because i wanted to check my knowledge base and other might be because of doubt
I will appreciate if I can get even basic answers for the questions. (i will dig up the rest on the basis of them).

waiting for further reply, and thanks again.

[Carlo Lazzaro]

Tarun:
what you're asking exceeds the goal of a forum like this one (in my opinion, at least).
If you laudably took this ambitious challenge for your master dissertation, the best advice is to study https://www.stata.com/bookstore/micr...metrics-stata/, which devotes chapters to what you're interested in, along with Stata codes and examples.

[Tarun Kumar]

thanks for the suggestion