[Mahadevan Suthakar]

Many thanks Carlo. One More clarification regarding unit root test,

'Unit root test usually relevant for T>N panel data set' - Is there any reference for the same to quote in my research paper ?

Kind regards,
Mahadevan

[Carlo Lazzaro]

Mahadevan:
https://www.stata.com/bookstore/micr...metrics-stata/ hosts a subparagraph (8.10.7; pages 278-279) on unit roots and cointegration within a paragraph (8.10; starting page: 271) that covers long panels.

[Mahadevan Suthakar]

Many thanks Carlo

Kind Regards,
Mahadevan

[Mahadevan Suthakar]

One clarification regarding Testing for time-fixed effects,

I Use the testparm (testparm _IYear_2012 _IYear_2013 _IYear_2014 _IYear_2015 _IYear_2016 _IYear_2017) command in stata after xtreg command (xi: xtreg ROAA SIZE CCAR LATA DPTA CRCO COIN REVD SLAR i.Year, fe)

The outcome is ( 1) _IYear_2012 = 0
( 2) _IYear_2013 = 0
( 3) _IYear_2014 = 0
( 4) _IYear_2015 = 0
( 5) _IYear_2016 = 0
( 6) _IYear_2017 = 0

F( 6, 46) = 2.24
Prob > F = 0.0556

Do we need to Apply dummy for time as per the above outcome?

What is the error level required to reject the null that all years coefficients are jointly equal to zero, so that time fixed effects are needed? Is there any reference which defines the required Prob value for testparm command?

Kind Regards,
Mahadevan

[Carlo Lazzaro]

Mahadevan:
if your Stata release supports -fvvarlist-, you can use the -i.year- notation and get rid of -xi:- prefix altogether.
As far as your second question is concerned the usual arbitrary 0.005 cut-off applies.

[Mahadevan Suthakar]

Sorry to trouble,

(a) Is it 5% (0.05) or 0.5% (0.005) ? Any reference available to the same ?
(b) If so, time dummy is not required for F( 6, 46) = 2.24 ; Prob > F = 0.0556 ?

Kind regards,
Mahadevan

[Carlo Lazzaro]

Mahadevan:
a) my mistake in previous reply: it's 5% (0.05) (no reference). Sorry for that.
b) correct, even though the p-value is borderline.

[Mahadevan Suthakar]

Many thanks Carlo. Great assistance for me to complete my study.

Kind regards,
Mahadevan